{"arm": "dim_no_gov", "item": "ROUTE-04", "family": "ROUTE", "rep": 0, "latency_s": 10.73, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"The sales team\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"D-701\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales team recorded the qualified_lead definition in SAL-POL-06.\"\n}\n```\n", "obj": {"answer": "The sales team", "owner_team": "sales", "source_id": "D-701", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales team recorded the qualified_lead definition in SAL-POL-06."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "AUTH-03", "family": "AUTH", "rep": 0, "latency_s": 12.44, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"318\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"N-201\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales-owned concept; support reported 402 but CP-ORG rule applies owner precedence.\"\n}\n```\n", "obj": {"answer": "318", "owner_team": "Sales", "source_id": "N-201", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Sales-owned concept; support reported 402 but CP-ORG rule applies owner precedence."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "STALE-05", "family": "STALE", "rep": 1, "latency_s": 13.59, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Yes, the Finance definition of logo churn (customers whose contract ended in the period and was not renewed within thirty days) is still in force.\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-301\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Valid since 2025-01-01 with no expiry date set\"\n}\n```\n", "obj": {"answer": "Yes, the Finance definition of logo churn (customers whose contract ended in the period and was not renewed within thirty days) is still in force.", "owner_team": "Finance", "source_id": "D-301", "policy_id": null, "unanswerable": false, "stale": false, "note": "Valid since 2025-01-01 with no expiry date set"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dimension", "item": "CONF-06", "family": "CONF", "rep": 0, "latency_s": 15.84, "context_chars": 11191, "raw": "```json\n{\"answer\": \"A person engaged under a services agreement with no employment relationship\", \"owner_team\": \"people\", \"source_id\": \"D-501\", \"policy_id\": \"CP-ORG\", \"unanswerable\": false, \"stale\": false, \"note\": \"People owns contractor concept; authoritative definition prevails per CP-ORG.\"}\n```\n", "obj": {"answer": "A person engaged under a services agreement with no employment relationship", "owner_team": "people", "source_id": "D-501", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "People owns contractor concept; authoritative definition prevails per CP-ORG."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "AUTH-02", "family": "AUTH", "rep": 1, "latency_s": 16.28, "context_chars": 6618, "raw": "```json\n{\"answer\": \"Sales, specifically the VP Sales\", \"owner_team\": \"Sales\", \"source_id\": null, \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"VP Sales owns active customer definition per organization records\"}\n```\n", "obj": {"answer": "Sales, specifically the VP Sales", "owner_team": "Sales", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "VP Sales owns active customer definition per organization records"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "GAP-03", "family": "GAP", "rep": 1, "latency_s": 10.74, "context_chars": 8932, "raw": "```json\n{\"answer\": null, \"owner_team\": null, \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"Gross margin definition not found in shared memory.\"}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Gross margin definition not found in shared memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "AUTH-01", "family": "AUTH", "rep": 1, "latency_s": 21.75, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"9.8m EUR\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"N-102\",\n  \"policy_id\": \"D-102\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance definition most current and restrictive; sales figure N-101 expired.\"\n}\n```\n", "obj": {"answer": "9.8m EUR", "owner_team": "finance", "source_id": "N-102", "policy_id": "D-102", "unanswerable": false, "stale": false, "note": "Finance definition most current and restrictive; sales figure N-101 expired."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "STALE-04", "family": "STALE", "rep": 0, "latency_s": 9.49, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"a lead with a named budget holder and a stated timeframe\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"D-701\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Valid from 2025-10-01 with no end date; authoritative sales definition.\"\n}\n```\n", "obj": {"answer": "a lead with a named budget holder and a stated timeframe", "owner_team": "sales", "source_id": "D-701", "policy_id": null, "unanswerable": false, "stale": false, "note": "Valid from 2025-10-01 with no end date; authoritative sales definition."}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "vector_memory", "item": "AUTH-02", "family": "AUTH", "rep": 1, "latency_s": 23.8, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Sales and support each own different definitions\",\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales (2024-09-01) defined as live contract; support (2025-02-01) as ticket activity\"\n}\n```\n", "obj": {"answer": "Sales and support each own different definitions", "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales (2024-09-01) defined as live contract; support (2025-02-01) as ticket activity"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "HIST-05", "family": "HIST", "rep": 1, "latency_s": 13.61, "context_chars": 6618, "raw": "```json\n{\"answer\": \"No\", \"owner_team\": \"Finance\", \"source_id\": \"FIN-POL-04\", \"policy_id\": null, \"unanswerable\": false, \"stale\": true, \"note\": \"Version 1 in force through 2026-03-31 lacked notice period exclusion\"}\n```\n", "obj": {"answer": "No", "owner_team": "Finance", "source_id": "FIN-POL-04", "policy_id": null, "unanswerable": false, "stale": true, "note": "Version 1 in force through 2026-03-31 lacked notice period exclusion"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "ROUTE-04", "family": "ROUTE", "rep": 0, "latency_s": 11.95, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"The VP Sales\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"D-701\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales team owns qualified_lead; VP Sales is the owner role\"\n}\n```\n", "obj": {"answer": "The VP Sales", "owner_team": "sales", "source_id": "D-701", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales team owns qualified_lead; VP Sales is the owner role"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "STALE-01", "family": "STALE", "rep": 1, "latency_s": 20.99, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"No. Legal, which owns this definition, updated the company position to five years effective 2026-06-01 (LEG-POL-04).\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"LEG-POL-04\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Support's seven-year definition follows Legal's superseded Version 1 policy.\"\n}\n```\n", "obj": {"answer": "No. Legal, which owns this definition, updated the company position to five years effective 2026-06-01 (LEG-POL-04).", "owner_team": "Legal", "source_id": "LEG-POL-04", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "Support's seven-year definition follows Legal's superseded Version 1 policy."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "vector_memory", "item": "HIST-02", "family": "HIST", "rep": 0, "latency_s": 15.47, "context_chars": 2694, "raw": "```json\n{\"answer\": \"seven years\", \"owner_team\": \"legal\", \"source_id\": null, \"policy_id\": null, \"unanswerable\": false, \"stale\": true, \"note\": \"Record from 2024-01-01 superseded by 2026-06-01 update to five years.\"}\n```\n", "obj": {"answer": "seven years", "owner_team": "legal", "source_id": null, "policy_id": null, "unanswerable": false, "stale": true, "note": "Record from 2024-01-01 superseded by 2026-06-01 update to five years."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "CONF-03", "family": "CONF", "rep": 0, "latency_s": 37.64, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Both stand. Engineering measures working days from confirmed order to despatch (D-601). Sales measures calendar days from PO to delivery on site (D-602). These are complementary measures of different process stages.\",\n  \"owner_team\": null,\n  \"source_id\": \"D-601, D-602\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Different definitions measure different process stages; both valid.\"\n}\n```\n", "obj": {"answer": "Both stand. Engineering measures working days from confirmed order to despatch (D-601). Sales measures calendar days from PO to delivery on site (D-602). These are complementary measures of different process stages.", "owner_team": null, "source_id": "D-601, D-602", "policy_id": null, "unanswerable": false, "stale": false, "note": "Different definitions measure different process stages; both valid."}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "STALE-02", "family": "STALE", "rep": 0, "latency_s": 23.51, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"No. The concept owner is support. Engineering's non-authoritative definition from 2025-11-01 was superseded by support's authoritative definition from 2026-01-15.\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"D-403\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Engineering recorded D-403 but support owns incident_severity_1. D-402 is current.\"\n}\n```\n", "obj": {"answer": "No. The concept owner is support. Engineering's non-authoritative definition from 2025-11-01 was superseded by support's authoritative definition from 2026-01-15.", "owner_team": "support", "source_id": "D-403", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "Engineering recorded D-403 but support owns incident_severity_1. D-402 is current."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "graph_catalog", "item": "GAP-04", "family": "GAP", "rep": 0, "latency_s": 7.66, "context_chars": 10621, "raw": "```json\n{\"answer\": null, \"owner_team\": null, \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"Net promoter score not found in memory\"}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Net promoter score not found in memory"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "GAP-01", "family": "GAP", "rep": 0, "latency_s": 10.08, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory contains 2026 ARR figures only, no 2027 target recorded.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains 2026 ARR figures only, no 2027 target recorded."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "graph_catalog", "item": "STALE-05", "family": "STALE", "rep": 0, "latency_s": 19.12, "context_chars": 10621, "raw": "```json\n{\"answer\": \"Yes. Definition: customers whose contract ended in the period and was not renewed within thirty days.\", \"owner_team\": \"Finance\", \"source_id\": \"D-301\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Valid from 2025-01-01, no end date\"}\n```\n", "obj": {"answer": "Yes. Definition: customers whose contract ended in the period and was not renewed within thirty days.", "owner_team": "Finance", "source_id": "D-301", "policy_id": null, "unanswerable": false, "stale": false, "note": "Valid from 2025-01-01, no end date"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "graph_catalog", "item": "HIST-01", "family": "HIST", "rep": 0, "latency_s": 23.22, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"committed subscription value recognised over the next twelve months, excluding one-off services\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-101 (Finance-owned) prevailed over competing D-103 (Sales) on 2026-02-01\"\n}\n```\n", "obj": {"answer": "committed subscription value recognised over the next twelve months, excluding one-off services", "owner_team": "Finance", "source_id": "D-101", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "D-101 (Finance-owned) prevailed over competing D-103 (Sales) on 2026-02-01"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "STALE-04", "family": "STALE", "rep": 1, "latency_s": 8.95, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"Yes, the Sales definition of qualified lead is in force: a lead with a named budget holder and a stated timeframe.\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"D-701\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Valid from 2025-10-01 with open end date. Authoritative.\"\n}\n```\n", "obj": {"answer": "Yes, the Sales definition of qualified lead is in force: a lead with a named budget holder and a stated timeframe.", "owner_team": "sales", "source_id": "D-701", "policy_id": null, "unanswerable": false, "stale": false, "note": "Valid from 2025-10-01 with open end date. Authoritative."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "graph_catalog", "item": "PROV-04", "family": "PROV", "rep": 0, "latency_s": 37.54, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"SUP-POL-03\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-402 (Support, owner) prevails over competing D-403 (Engineering) per CP-ORG\"\n}\n```\n", "obj": {"answer": "SUP-POL-03", "owner_team": "Support", "source_id": "D-402", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "D-402 (Support, owner) prevails over competing D-403 (Engineering) per CP-ORG"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "CONF-04", "family": "CONF", "rep": 1, "latency_s": 41.28, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"5 years after the contract ends\",\n  \"owner_team\": \"legal\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal updated to 5 years (2026-06-01); Support shows 7 years same date. Legal is authoritative.\"\n}\n```\n", "obj": {"answer": "5 years after the contract ends", "owner_team": "legal", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal updated to 5 years (2026-06-01); Support shows 7 years same date. Legal is authoritative."}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 1, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "CONF-05", "family": "CONF", "rep": 1, "latency_s": 15.52, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory records different definitions from Support and Engineering but does not state which applies company-wide.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory records different definitions from Support and Engineering but does not state which applies company-wide."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "CONF-01", "family": "CONF", "rep": 1, "latency_s": 30.85, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Finance's 9.8m EUR (FIN-STMT-Q1) stands. Rule CP-ORG: the record written by the concept owner prevails.\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"FIN-STMT-Q1\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance owns ARR definition; Sales record retained as Sales' view.\"\n}\n```\n", "obj": {"answer": "Finance's 9.8m EUR (FIN-STMT-Q1) stands. Rule CP-ORG: the record written by the concept owner prevails.", "owner_team": "Finance", "source_id": "FIN-STMT-Q1", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Finance owns ARR definition; Sales record retained as Sales' view."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "HIST-04", "family": "HIST", "rep": 1, "latency_s": 49.79, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"version 2\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Engineering recorded conflicting version 1 (D-403); support version 2 more recent.\"\n}\n```\n", "obj": {"answer": "version 2", "owner_team": "support", "source_id": "D-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "Engineering recorded conflicting version 1 (D-403); support version 2 more recent."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "GAP-02", "family": "GAP", "rep": 1, "latency_s": 7.72, "context_chars": 6618, "raw": "```json\n{\"answer\": null, \"owner_team\": \"People\", \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"No total headcount recorded in shared memory.\"}\n```\n", "obj": {"answer": null, "owner_team": "People", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "No total headcount recorded in shared memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "wiki", "item": "GAP-02", "family": "GAP", "rep": 0, "latency_s": 14.82, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": \"People\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"No total headcount recorded in shared memory.\"\n}\n```\n", "obj": {"answer": null, "owner_team": "People", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "No total headcount recorded in shared memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "bitemporal", "item": "ROUTE-03", "family": "ROUTE", "rep": 1, "latency_s": 18.23, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Ask the Legal team\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal owns data retention policy definitions\"\n}\n```\n", "obj": {"answer": "Ask the Legal team", "owner_team": "Legal", "source_id": "D-802", "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal owns data retention policy definitions"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "STALE-02", "family": "STALE", "rep": 1, "latency_s": 29.63, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"No. Support owns the severity 1 incident definition. The company uses Support's definition from 2026-01-15: an outage affecting any customer with no workaround, or any confirmed data loss.\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"SUP-POL-03\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support owns concept; their 2026-01-15 definition prevails per CP-ORG.\"\n}\n```\n", "obj": {"answer": "No. Support owns the severity 1 incident definition. The company uses Support's definition from 2026-01-15: an outage affecting any customer with no workaround, or any confirmed data loss.", "owner_team": "Support", "source_id": "SUP-POL-03", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Support owns concept; their 2026-01-15 definition prevails per CP-ORG."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "graph_catalog", "item": "AUTH-04", "family": "AUTH", "rep": 0, "latency_s": 12.43, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"General Counsel\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"LEG-POL-04\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Concept owned by Legal per governance catalogue.\"\n}\n```\n", "obj": {"answer": "General Counsel", "owner_team": "Legal", "source_id": "LEG-POL-04", "policy_id": null, "unanswerable": false, "stale": false, "note": "Concept owned by Legal per governance catalogue."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "ROUTE-01", "family": "ROUTE", "rep": 1, "latency_s": 17.42, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"The Head of Finance must agree; finance team owns the annual recurring revenue concept\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-102\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance owns ARR concept; their definitions prevail by policy\"\n}\n```\n", "obj": {"answer": "The Head of Finance must agree; finance team owns the annual recurring revenue concept", "owner_team": "finance", "source_id": "D-102", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance owns ARR concept; their definitions prevail by policy"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "GAP-01", "family": "GAP", "rep": 1, "latency_s": 13.0, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": \"finance\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory contains 2026 ARR figures only; no 2027 target documented\"\n}\n```\n", "obj": {"answer": null, "owner_team": "finance", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains 2026 ARR figures only; no 2027 target documented"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "bitemporal", "item": "ROUTE-01", "family": "ROUTE", "rep": 0, "latency_s": 52.14, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Finance and Sales must both agree; they have conflicting current ARR definitions\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-102\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance (D-102) most recent; Sales also has current definition (D-103)\"\n}\n```\n", "obj": {"answer": "Finance and Sales must both agree; they have conflicting current ARR definitions", "owner_team": "Finance", "source_id": "D-102", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance (D-102) most recent; Sales also has current definition (D-103)"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "AUTH-05", "family": "AUTH", "rep": 1, "latency_s": 40.75, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Both 23 working days (ENG) and 41 calendar days (SALES) valid on 2026-06-01 with different definitions\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Both 23 working days (ENG) and 41 calendar days (SALES) valid on 2026-06-01 with different definitions"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "HIST-05", "family": "HIST", "rep": 0, "latency_s": 16.9, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"No\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-101 valid 2025-12-31, excluded one-off services only. D-102 introduced notice period exclusion 2026-04-01.\"\n}\n```\n", "obj": {"answer": "No", "owner_team": "finance", "source_id": "D-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "D-101 valid 2025-12-31, excluded one-off services only. D-102 introduced notice period exclusion 2026-04-01."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "GAP-03", "family": "GAP", "rep": 1, "latency_s": 13.53, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Gross margin is not defined in company records\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Gross margin is not defined in company records"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dimension", "item": "CONF-02", "family": "CONF", "rep": 1, "latency_s": 22.47, "context_chars": 11191, "raw": "```json\n{\"answer\": \"318 (Sales)\", \"owner_team\": \"sales\", \"source_id\": \"N-201\", \"policy_id\": \"CP-ORG\", \"unanswerable\": false, \"stale\": false, \"note\": \"Sales owns active_customer concept. N-201 authoritative; N-202 is support's non-authoritative view.\"}\n```\n", "obj": {"answer": "318 (Sales)", "owner_team": "sales", "source_id": "N-201", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Sales owns active_customer concept. N-201 authoritative; N-202 is support's non-authoritative view."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "STALE-02", "family": "STALE", "rep": 0, "latency_s": 49.12, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"No\",\n  \"owner_team\": \"engineering\",\n  \"source_id\": \"D-403\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-402 (support, 2026-01-15) superseded D-403 as current definition\"\n}\n```\n", "obj": {"answer": "No", "owner_team": "engineering", "source_id": "D-403", "policy_id": null, "unanswerable": false, "stale": true, "note": "D-402 (support, 2026-01-15) superseded D-403 as current definition"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "vector_memory", "item": "GAP-04", "family": "GAP", "rep": 1, "latency_s": 7.76, "context_chars": 2694, "raw": "```json\n{\"answer\": null, \"owner_team\": null, \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"Net promoter score definition not found in shared memory.\"}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Net promoter score definition not found in shared memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "graph_catalog", "item": "PROV-01", "family": "PROV", "rep": 1, "latency_s": 17.95, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"FIN-POL-07\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-102\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance-owned definition (D-102, valid 2026-04-01) prevails over Sales' concurrent definition D-103 per CP-ORG rule.\"\n}\n```\n", "obj": {"answer": "FIN-POL-07", "owner_team": "Finance", "source_id": "D-102", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Finance-owned definition (D-102, valid 2026-04-01) prevails over Sales' concurrent definition D-103 per CP-ORG rule."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "PROV-04", "family": "PROV", "rep": 1, "latency_s": 37.29, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"SUP-POL-03\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-402 (authoritative, support) prevails over D-403 (engineering, non-authoritative)\"\n}\n```\n", "obj": {"answer": "SUP-POL-03", "owner_team": "support", "source_id": "D-402", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "D-402 (authoritative, support) prevails over D-403 (engineering, non-authoritative)"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "CONF-01", "family": "CONF", "rep": 0, "latency_s": 63.55, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Finance figure of 9.8m EUR (N-102) stands under rule D-102 (committed subscription value over 12 months, excluding one-off services and customers in notice).\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"N-102\",\n  \"policy_id\": \"D-102\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Sales figure N-101 expired 2026-04-30. D-102 is current ARR definition.\"\n}\n```\n", "obj": {"answer": "Finance figure of 9.8m EUR (N-102) stands under rule D-102 (committed subscription value over 12 months, excluding one-off services and customers in notice).", "owner_team": "finance", "source_id": "N-102", "policy_id": "D-102", "unanswerable": false, "stale": true, "note": "Sales figure N-101 expired 2026-04-30. D-102 is current ARR definition."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "AUTH-06", "family": "AUTH", "rep": 0, "latency_s": 15.18, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"People\",\n  \"owner_team\": \"People\",\n  \"source_id\": \"D-501\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"People team owns contractor concept; Head of People accountable.\"\n}\n```\n", "obj": {"answer": "People", "owner_team": "People", "source_id": "D-501", "policy_id": null, "unanswerable": false, "stale": false, "note": "People team owns contractor concept; Head of People accountable."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "CONF-01", "family": "CONF", "rep": 1, "latency_s": 14.88, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"Finance's 9.8m EUR (N-102) stands. Finance owns the ARR concept, so per CP-ORG, the finance-recorded figure prevails over sales' conflicting 12.4m EUR figure (N-101, now stale).\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"N-102\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales' N-101 expired 2026-04-30; finance's N-102 current.\"\n}\n```\n", "obj": {"answer": "Finance's 9.8m EUR (N-102) stands. Finance owns the ARR concept, so per CP-ORG, the finance-recorded figure prevails over sales' conflicting 12.4m EUR figure (N-101, now stale).", "owner_team": "finance", "source_id": "N-102", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Sales' N-101 expired 2026-04-30; finance's N-102 current."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "PROV-03", "family": "PROV", "rep": 1, "latency_s": 50.32, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Legal policy; 5 years (effective 2026-06-01)\",\n  \"owner_team\": \"legal\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Memory lacks document names. Support rule conflicts at 7 years; legal supersedes.\"\n}\n```\n", "obj": {"answer": "Legal policy; 5 years (effective 2026-06-01)", "owner_team": "legal", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Memory lacks document names. Support rule conflicts at 7 years; legal supersedes."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "CONF-04", "family": "CONF", "rep": 1, "latency_s": 35.23, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Five years after contract ends (Legal policy D-802). Support records seven years (D-803), which conflicts.\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Conflict: Legal D-802 (5yr) vs Support D-803 (7yr). Legal is authoritative on compliance.\"\n}\n```\n", "obj": {"answer": "Five years after contract ends (Legal policy D-802). Support records seven years (D-803), which conflicts.", "owner_team": "legal", "source_id": "D-802", "policy_id": null, "unanswerable": false, "stale": false, "note": "Conflict: Legal D-802 (5yr) vs Support D-803 (7yr). Legal is authoritative on compliance."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "GAP-02", "family": "GAP", "rep": 1, "latency_s": 9.21, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": \"people\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory records headcount growth plan (engineering +11 in 2027) but not total current headcount.\"\n}\n```\n", "obj": {"answer": null, "owner_team": "people", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory records headcount growth plan (engineering +11 in 2027) but not total current headcount."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "vector_memory", "item": "ROUTE-02", "family": "ROUTE", "rep": 1, "latency_s": 9.12, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Support's decision. Support owns the current definition of incident_severity_1 as of 2026-01-15.\",\n  \"owner_team\": \"support\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support defined incident_severity_1 on 2026-01-15, establishing ownership.\"\n}\n```\n", "obj": {"answer": "Support's decision. Support owns the current definition of incident_severity_1 as of 2026-01-15.", "owner_team": "support", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Support defined incident_severity_1 on 2026-01-15, establishing ownership."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "STALE-02", "family": "STALE", "rep": 1, "latency_s": 53.69, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"No\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Support's 2026-01-15 definition (D-402) supersedes Engineering's 2025-11-01 definition (D-403).\"\n}\n```\n", "obj": {"answer": "No", "owner_team": "Support", "source_id": "D-402", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "Support's 2026-01-15 definition (D-402) supersedes Engineering's 2025-11-01 definition (D-403)."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "bitemporal", "item": "PROV-01", "family": "PROV", "rep": 1, "latency_s": 26.32, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"FIN-POL-07\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-102\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-102 valid from 2026-04-01 onwards; superseded D-101\"\n}\n```\n", "obj": {"answer": "FIN-POL-07", "owner_team": "Finance", "source_id": "D-102", "policy_id": null, "unanswerable": false, "stale": false, "note": "D-102 valid from 2026-04-01 onwards; superseded D-101"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "CONF-05", "family": "CONF", "rep": 0, "latency_s": 10.39, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Support's definition applies: an outage affecting any customer with no workaround, or any confirmed data loss (Version 2, in force 2026-01-15, still current).\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"SUP-POL-03\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support owns severity 1 incident definition. Engineering's record explicitly defers to Support.\"\n}\n```\n", "obj": {"answer": "Support's definition applies: an outage affecting any customer with no workaround, or any confirmed data loss (Version 2, in force 2026-01-15, still current).", "owner_team": "Support", "source_id": "SUP-POL-03", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Support owns severity 1 incident definition. Engineering's record explicitly defers to Support."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "ROUTE-01", "family": "ROUTE", "rep": 0, "latency_s": 38.48, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Finance and Sales teams must agree; they define ARR differently\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"2026-04-01\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Two teams maintain different ARR definitions; both must approve change\"\n}\n```\n", "obj": {"answer": "Finance and Sales teams must agree; they define ARR differently", "owner_team": "finance", "source_id": "2026-04-01", "policy_id": null, "unanswerable": false, "stale": false, "note": "Two teams maintain different ARR definitions; both must approve change"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "GAP-03", "family": "GAP", "rep": 0, "latency_s": 8.71, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"No gross margin definition found in organizational memory.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "No gross margin definition found in organizational memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "graph_catalog", "item": "STALE-05", "family": "STALE", "rep": 1, "latency_s": 10.64, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Yes. Finance definition valid since 2025-01-01 with no end date: customers whose contract ended in the period and was not renewed within thirty days.\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-301\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Definition has open validity period; still in force.\"\n}\n```\n", "obj": {"answer": "Yes. Finance definition valid since 2025-01-01 with no end date: customers whose contract ended in the period and was not renewed within thirty days.", "owner_team": "Finance", "source_id": "D-301", "policy_id": null, "unanswerable": false, "stale": false, "note": "Definition has open validity period; still in force."}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "graph_catalog", "item": "STALE-03", "family": "STALE", "rep": 1, "latency_s": 11.71, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"No. Version 1 (D-101) expired on 2026-03-31 and is superseded by D-102 (valid from 2026-04-01).\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-101 invalid as of 2026-03-31; D-102 currently in force\"\n}\n```\n", "obj": {"answer": "No. Version 1 (D-101) expired on 2026-03-31 and is superseded by D-102 (valid from 2026-04-01).", "owner_team": "Finance", "source_id": "D-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "D-101 invalid as of 2026-03-31; D-102 currently in force"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dim_no_gov", "item": "HIST-02", "family": "HIST", "rep": 1, "latency_s": 17.2, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"customer records are kept for seven years after the contract ends\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-801\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Valid on 2026-03-01; superseded by D-802 from 2026-06-01\"\n}\n```\n", "obj": {"answer": "customer records are kept for seven years after the contract ends", "owner_team": "legal", "source_id": "D-801", "policy_id": null, "unanswerable": false, "stale": true, "note": "Valid on 2026-03-01; superseded by D-802 from 2026-06-01"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "STALE-03", "family": "STALE", "rep": 0, "latency_s": 28.57, "context_chars": 6618, "raw": "```json\n{\"answer\": \"No. Finance version 1 was in force until 2026-03-31 and has been superseded by version 2 from 2026-04-01.\", \"owner_team\": \"Finance\", \"source_id\": \"FIN-POL-04\", \"policy_id\": null, \"unanswerable\": false, \"stale\": true, \"note\": \"Version 1 ended; version 2 now current\"}\n```\n", "obj": {"answer": "No. Finance version 1 was in force until 2026-03-31 and has been superseded by version 2 from 2026-04-01.", "owner_team": "Finance", "source_id": "FIN-POL-04", "policy_id": null, "unanswerable": false, "stale": true, "note": "Version 1 ended; version 2 now current"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dimension", "item": "CONF-05", "family": "CONF", "rep": 1, "latency_s": 15.86, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"an outage affecting any customer with no workaround, or any confirmed data loss\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support owns the concept; their authoritative definition D-402 (from 2026-01-15) supersedes Engineering's non-authoritative D-403.\"\n}\n```\n", "obj": {"answer": "an outage affecting any customer with no workaround, or any confirmed data loss", "owner_team": "support", "source_id": "D-402", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Support owns the concept; their authoritative definition D-402 (from 2026-01-15) supersedes Engineering's non-authoritative D-403."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "STALE-05", "family": "STALE", "rep": 0, "latency_s": 12.6, "context_chars": 11191, "raw": "```json\n{\"answer\": \"Yes, the Finance definition of logo churn is still in force. Valid period runs from 2025-01-01 with no end date.\", \"owner_team\": \"finance\", \"source_id\": \"D-301\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Definition remains active; valid_to is open as of 2026-09-07.\"}\n```\n", "obj": {"answer": "Yes, the Finance definition of logo churn is still in force. Valid period runs from 2025-01-01 with no end date.", "owner_team": "finance", "source_id": "D-301", "policy_id": null, "unanswerable": false, "stale": false, "note": "Definition remains active; valid_to is open as of 2026-09-07."}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "vector_memory", "item": "CONF-02", "family": "CONF", "rep": 0, "latency_s": 13.48, "context_chars": 2694, "raw": "```json\n{\"answer\": null, \"owner_team\": null, \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"Two teams define active_customer differently; memory lacks precedence rule\"}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Two teams define active_customer differently; memory lacks precedence rule"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "PROV-03", "family": "PROV", "rep": 0, "latency_s": 39.95, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Legal record (2026-06-01): customer records kept five years after contract ends\",\n  \"owner_team\": \"legal\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Supersedes 2024 legal rule; Support records conflicting 7-year requirement (same date)\"\n}\n```\n", "obj": {"answer": "Legal record (2026-06-01): customer records kept five years after contract ends", "owner_team": "legal", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Supersedes 2024 legal rule; Support records conflicting 7-year requirement (same date)"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "HIST-03", "family": "HIST", "rep": 0, "latency_s": 62.44, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"1\",\n  \"owner_team\": \"support, engineering\",\n  \"source_id\": \"D-401, D-403\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Two v1 definitions both active on 2025-12-01; D-401 later superseded\"\n}\n```\n", "obj": {"answer": "1", "owner_team": "support, engineering", "source_id": "D-401, D-403", "policy_id": null, "unanswerable": false, "stale": true, "note": "Two v1 definitions both active on 2025-12-01; D-401 later superseded"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "ROUTE-02", "family": "ROUTE", "rep": 1, "latency_s": 32.88, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"Support Lead, as the concept owner for severity 1 incidents\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Owners control concept definition per conflict policy\"\n}\n```\n", "obj": {"answer": "Support Lead, as the concept owner for severity 1 incidents", "owner_team": "support", "source_id": "D-402", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Owners control concept definition per conflict policy"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "GAP-01", "family": "GAP", "rep": 0, "latency_s": 11.29, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": \"Finance\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory contains 2026 ARR figures only, no 2027 target documented\"\n}\n```\n", "obj": {"answer": null, "owner_team": "Finance", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains 2026 ARR figures only, no 2027 target documented"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dimension", "item": "PROV-04", "family": "PROV", "rep": 0, "latency_s": 11.8, "context_chars": 11191, "raw": "```json\n{\"answer\": \"SUP-POL-03 (Support Policy 03)\", \"owner_team\": \"support\", \"source_id\": \"D-402\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"D-402 is authoritative, valid from 2026-01-15 to open\"}\n```\n", "obj": {"answer": "SUP-POL-03 (Support Policy 03)", "owner_team": "support", "source_id": "D-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "D-402 is authoritative, valid from 2026-01-15 to open"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "CONF-02", "family": "CONF", "rep": 1, "latency_s": 21.0, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": \"N-201, N-202\",\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Sales and Support use conflicting active_customer definitions; both currently valid, figures not comparable.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": "N-201, N-202", "policy_id": null, "unanswerable": true, "stale": false, "note": "Sales and Support use conflicting active_customer definitions; both currently valid, figures not comparable."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "AUTH-02", "family": "AUTH", "rep": 0, "latency_s": 16.13, "context_chars": 11191, "raw": "```json\n{\"answer\": \"The sales team, specifically the VP Sales role\", \"owner_team\": \"sales\", \"source_id\": \"D-201\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Concept ownership confirmed by authoritative definition D-201 from sales.\"}\n```\n", "obj": {"answer": "The sales team, specifically the VP Sales role", "owner_team": "sales", "source_id": "D-201", "policy_id": null, "unanswerable": false, "stale": false, "note": "Concept ownership confirmed by authoritative definition D-201 from sales."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "HIST-05", "family": "HIST", "rep": 0, "latency_s": 9.69, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"No\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"FIN-POL-04\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Version 1 definition (in force through 2026-03-31) did not exclude notice period customers.\"\n}\n```\n", "obj": {"answer": "No", "owner_team": "Finance", "source_id": "FIN-POL-04", "policy_id": null, "unanswerable": false, "stale": true, "note": "Version 1 definition (in force through 2026-03-31) did not exclude notice period customers."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "CONF-04", "family": "CONF", "rep": 0, "latency_s": 37.71, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"5 years after the contract ends\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal D-802 authoritative; Support D-803 conflicts stating 7 years\"\n}\n```\n", "obj": {"answer": "5 years after the contract ends", "owner_team": "Legal", "source_id": "D-802", "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal D-802 authoritative; Support D-803 conflicts stating 7 years"}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 1, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "CONF-06", "family": "CONF", "rep": 1, "latency_s": 13.22, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"any person invoicing the company who is not on payroll\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-502\",\n  \"policy_id\": \"D-502\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance's 2025-08-01 definition (D-502) supersedes People's 2024-04-01 definition (D-501)\"\n}\n```\n", "obj": {"answer": "any person invoicing the company who is not on payroll", "owner_team": "Finance", "source_id": "D-502", "policy_id": "D-502", "unanswerable": false, "stale": false, "note": "Finance's 2025-08-01 definition (D-502) supersedes People's 2024-04-01 definition (D-501)"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 1, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "STALE-04", "family": "STALE", "rep": 0, "latency_s": 14.04, "context_chars": 8932, "raw": "```json\n{\"answer\": \"Yes, the Sales definition of a qualified lead remains in force: 'a lead with a named budget holder and a stated timeframe'\", \"owner_team\": \"sales\", \"source_id\": \"D-701\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Valid from 2025-10-01 with no end date.\"}\n```\n", "obj": {"answer": "Yes, the Sales definition of a qualified lead remains in force: 'a lead with a named budget holder and a stated timeframe'", "owner_team": "sales", "source_id": "D-701", "policy_id": null, "unanswerable": false, "stale": false, "note": "Valid from 2025-10-01 with no end date."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "graph_catalog", "item": "AUTH-03", "family": "AUTH", "rep": 1, "latency_s": 11.17, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"318\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"N-201\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales owns active_customer; their 318 figure prevails over Support's 402.\"\n}\n```\n", "obj": {"answer": "318", "owner_team": "Sales", "source_id": "N-201", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Sales owns active_customer; their 318 figure prevails over Support's 402."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "PROV-02", "family": "PROV", "rep": 0, "latency_s": 12.79, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Sales\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"N-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Report from SAL-DECK-Q1 expired 2026-04-30. Finance now reports 9.8m EUR.\"\n}\n```\n", "obj": {"answer": "Sales", "owner_team": "Finance", "source_id": "N-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "Report from SAL-DECK-Q1 expired 2026-04-30. Finance now reports 9.8m EUR."}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "AUTH-01", "family": "AUTH", "rep": 1, "latency_s": 18.12, "context_chars": 11191, "raw": "```json\n{\"answer\": \"Finance team; 9.8m EUR\", \"owner_team\": \"finance\", \"source_id\": \"N-102\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Authoritative current figure from concept owner (finance). Sales figure N-101 expired 2026-04-30.\"}\n```\n", "obj": {"answer": "Finance team; 9.8m EUR", "owner_team": "finance", "source_id": "N-102", "policy_id": null, "unanswerable": false, "stale": false, "note": "Authoritative current figure from concept owner (finance). Sales figure N-101 expired 2026-04-30."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "PROV-01", "family": "PROV", "rep": 0, "latency_s": 46.5, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Finance record dated 2026-04-01\",\n  \"owner_team\": \"finance\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Current finance definition (2026-04-01) supersedes earlier sales definition from 2025-06-01\"\n}\n```\n", "obj": {"answer": "Finance record dated 2026-04-01", "owner_team": "finance", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Current finance definition (2026-04-01) supersedes earlier sales definition from 2025-06-01"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "AUTH-03", "family": "AUTH", "rep": 0, "latency_s": 10.45, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"318\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"SAL-REP-JUL\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales owns the definition; their count prevails over Support's 402.\"\n}\n```\n", "obj": {"answer": "318", "owner_team": "Sales", "source_id": "SAL-REP-JUL", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Sales owns the definition; their count prevails over Support's 402."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "CONF-02", "family": "CONF", "rep": 0, "latency_s": 12.59, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"Sales' figure of 318 stands\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"N-201\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales owns active_customer concept. N-201 (sales, authoritative) prevails over N-202 (support, non-authoritative) per conflict policy.\"\n}\n```\n", "obj": {"answer": "Sales' figure of 318 stands", "owner_team": "sales", "source_id": "N-201", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Sales owns active_customer concept. N-201 (sales, authoritative) prevails over N-202 (support, non-authoritative) per conflict policy."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "ROUTE-05", "family": "ROUTE", "rep": 0, "latency_s": 14.93, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"Engineering Manager\",\n  \"owner_team\": \"engineering\",\n  \"source_id\": \"D-601\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Authoritative definition recorded by engineering from 2025-05-01, valid to open.\"\n}\n```\n", "obj": {"answer": "Engineering Manager", "owner_team": "engineering", "source_id": "D-601", "policy_id": null, "unanswerable": false, "stale": false, "note": "Authoritative definition recorded by engineering from 2025-05-01, valid to open."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "PROV-01", "family": "PROV", "rep": 0, "latency_s": 61.85, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"FIN-POL-07\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-102\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Most recent definition, valid from 2026-04-01\"\n}\n```\n", "obj": {"answer": "FIN-POL-07", "owner_team": "Finance", "source_id": "D-102", "policy_id": null, "unanswerable": false, "stale": false, "note": "Most recent definition, valid from 2026-04-01"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "AUTH-06", "family": "AUTH", "rep": 0, "latency_s": 20.91, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Finance: any person invoicing the company who is not on payroll\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"2025-08-01\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance definition is most recent; People team provided earlier definition\"\n}\n```\n", "obj": {"answer": "Finance: any person invoicing the company who is not on payroll", "owner_team": "finance", "source_id": "2025-08-01", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance definition is most recent; People team provided earlier definition"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "AUTH-01", "family": "AUTH", "rep": 0, "latency_s": 11.88, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"9.8m EUR\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"N-102\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance owns ARR concept; their current figure prevails over Sales' expired Q1 figure.\"\n}\n```\n", "obj": {"answer": "9.8m EUR", "owner_team": "Finance", "source_id": "N-102", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Finance owns ARR concept; their current figure prevails over Sales' expired Q1 figure."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "GAP-02", "family": "GAP", "rep": 1, "latency_s": 11.13, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory contains only headcount growth plan for engineering in 2027, not total headcount.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains only headcount growth plan for engineering in 2027, not total headcount."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dimension", "item": "AUTH-02", "family": "AUTH", "rep": 1, "latency_s": 12.17, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"sales\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"D-201\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales team (VP Sales) owns active_customer definition per concept ownership record.\"\n}\n```\n", "obj": {"answer": "sales", "owner_team": "sales", "source_id": "D-201", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales team (VP Sales) owns active_customer definition per concept ownership record."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "STALE-04", "family": "STALE", "rep": 1, "latency_s": 12.51, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Yes, still in force as of 2026-09-07.\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"SAL-POL-06\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Version 1 effective from 2025-10-01, marked still current.\"\n}\n```\n", "obj": {"answer": "Yes, still in force as of 2026-09-07.", "owner_team": "Sales", "source_id": "SAL-POL-06", "policy_id": null, "unanswerable": false, "stale": false, "note": "Version 1 effective from 2025-10-01, marked still current."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dim_no_gov", "item": "ROUTE-01", "family": "ROUTE", "rep": 1, "latency_s": 33.3, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Finance and sales must both agree, as they have conflicting current definitions of ARR. Finance defines it in policy D-102 (forward-looking subscription value over next twelve months, excluding one-off services and customers in notice). Sales defines it in wiki entry D-103 (total contracts signed in last twelve months, including one-off services).\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-102\",\n  \"policy_id\": \"FIN-POL-07\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  ", "obj": {"answer": "Finance and sales must both agree, as they have conflicting current definitions of ARR. Finance defines it in policy D-102 (forward-looking subscription value over next twelve months, excluding one-off services and customers in notice). Sales defines it in wiki entry D-103 (total contracts signed in last twelve months, including one-off services).", "owner_team": "finance", "source_id": "D-102", "policy_id": "FIN-POL-07", "unanswerable": false, "stale": false, "note": "Two teams have materially different ARR definitions currently active"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "ROUTE-03", "family": "ROUTE", "rep": 0, "latency_s": 22.64, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"legal\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal team owns data retention policy; support has conflicting definition\"\n}\n```\n", "obj": {"answer": "legal", "owner_team": "legal", "source_id": "D-802", "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal team owns data retention policy; support has conflicting definition"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "STALE-05", "family": "STALE", "rep": 1, "latency_s": 11.17, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"Yes. Finance definition D-301 (FIN-POL-05) defines logo churn as 'customers whose contract ended in the period and was not renewed within thirty days', valid from 2025-01-01 with no end date.\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-301\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Authoritative definition remains in force as of 2026-09-07\"\n}\n```\n", "obj": {"answer": "Yes. Finance definition D-301 (FIN-POL-05) defines logo churn as 'customers whose contract ended in the period and was not renewed within thirty days', valid from 2025-01-01 with no end date.", "owner_team": "finance", "source_id": "D-301", "policy_id": null, "unanswerable": false, "stale": false, "note": "Authoritative definition remains in force as of 2026-09-07"}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "bitemporal", "item": "AUTH-04", "family": "AUTH", "rep": 1, "latency_s": 23.16, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Legal\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal wrote policy definition; Support wiki conflicts on duration\"\n}\n```\n", "obj": {"answer": "Legal", "owner_team": "Legal", "source_id": "D-802", "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal wrote policy definition; Support wiki conflicts on duration"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "ROUTE-01", "family": "ROUTE", "rep": 1, "latency_s": 54.63, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Finance and Sales, who have conflicting ARR definitions (D-102 and D-103) both currently in force\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-102, D-103\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance's D-102 and Sales's D-103 describe ARR differently\"\n}\n```\n", "obj": {"answer": "Finance and Sales, who have conflicting ARR definitions (D-102 and D-103) both currently in force", "owner_team": "Finance", "source_id": "D-102, D-103", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance's D-102 and Sales's D-103 describe ARR differently"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "STALE-01", "family": "STALE", "rep": 1, "latency_s": 18.05, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"No. The current company position is five years after contract ends (Legal policy D-802, 2026-06-01). Support wiki's seven-year position (D-803) is superseded by the owning team's authoritative standard.\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Legal owns data_retention; their policy supersedes non-owner views.\"\n}\n```\n", "obj": {"answer": "No. The current company position is five years after contract ends (Legal policy D-802, 2026-06-01). Support wiki's seven-year position (D-803) is superseded by the owning team's authoritative standard.", "owner_team": "Legal", "source_id": "D-802", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "Legal owns data_retention; their policy supersedes non-owner views."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dimension", "item": "GAP-04", "family": "GAP", "rep": 1, "latency_s": 10.41, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"No record of net promoter score in organizational memory.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "No record of net promoter score in organizational memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "wiki", "item": "STALE-05", "family": "STALE", "rep": 1, "latency_s": 8.44, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Yes\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"FIN-POL-05\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Version 1 in force from 2025-01-01, still current\"\n}\n```\n", "obj": {"answer": "Yes", "owner_team": "Finance", "source_id": "FIN-POL-05", "policy_id": null, "unanswerable": false, "stale": false, "note": "Version 1 in force from 2025-01-01, still current"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "wiki", "item": "AUTH-06", "family": "AUTH", "rep": 1, "latency_s": 16.75, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Head of People\",\n  \"owner_team\": \"People\",\n  \"source_id\": \"PPL-POL-02\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"People team owns contractor definition, in force 2024-04-01, still current.\"\n}\n```\n", "obj": {"answer": "Head of People", "owner_team": "People", "source_id": "PPL-POL-02", "policy_id": null, "unanswerable": false, "stale": false, "note": "People team owns contractor definition, in force 2024-04-01, still current."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "AUTH-04", "family": "AUTH", "rep": 1, "latency_s": 19.98, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"The General Counsel (Legal)\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal owns the definition per organization governance structure.\"\n}\n```\n", "obj": {"answer": "The General Counsel (Legal)", "owner_team": "Legal", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal owns the definition per organization governance structure."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "PROV-04", "family": "PROV", "rep": 1, "latency_s": 13.37, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": \"support\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory defines severity 1 (2026-01-15 support) but not document location.\"\n}\n```\n", "obj": {"answer": null, "owner_team": "support", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory defines severity 1 (2026-01-15 support) but not document location."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "STALE-01", "family": "STALE", "rep": 0, "latency_s": 16.86, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"No. Legal's current policy (D-802, from 2026-06-01) specifies five years. Support's seven-year statement (D-803, same date) is superseded per CP-ORG.\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Support's definition contradicts Legal's (owner) position; Legal's prevails.\"\n}\n```\n", "obj": {"answer": "No. Legal's current policy (D-802, from 2026-06-01) specifies five years. Support's seven-year statement (D-803, same date) is superseded per CP-ORG.", "owner_team": "Legal", "source_id": "D-802", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "Support's definition contradicts Legal's (owner) position; Legal's prevails."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "wiki", "item": "ROUTE-01", "family": "ROUTE", "rep": 1, "latency_s": 29.87, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"The Head of Finance\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"FIN-POL-07\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance owns the definition of annual recurring revenue.\"\n}\n```\n", "obj": {"answer": "The Head of Finance", "owner_team": "Finance", "source_id": "FIN-POL-07", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance owns the definition of annual recurring revenue."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "HIST-01", "family": "HIST", "rep": 1, "latency_s": 30.0, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"committed subscription value recognised over the next twelve months, excluding one-off services\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-103 from Sales provided conflicting definition from June 2025.\"\n}\n```\n", "obj": {"answer": "committed subscription value recognised over the next twelve months, excluding one-off services", "owner_team": "Finance", "source_id": "D-101", "policy_id": null, "unanswerable": false, "stale": false, "note": "D-103 from Sales provided conflicting definition from June 2025."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "GAP-03", "family": "GAP", "rep": 0, "latency_s": 8.35, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"No gross margin definition found in shared memory.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "No gross margin definition found in shared memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "vector_memory", "item": "CONF-03", "family": "CONF", "rep": 0, "latency_s": 18.21, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory contains both definitions with no hierarchy or resolution rule.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains both definitions with no hierarchy or resolution rule."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "AUTH-05", "family": "AUTH", "rep": 1, "latency_s": 26.5, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"23 working days\",\n  \"owner_team\": \"Engineering\",\n  \"source_id\": \"N-401\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Engineering owns concept; N-401 prevails per CP-ORG governance rule\"\n}\n```\n", "obj": {"answer": "23 working days", "owner_team": "Engineering", "source_id": "N-401", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Engineering owns concept; N-401 prevails per CP-ORG governance rule"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "CONF-03", "family": "CONF", "rep": 1, "latency_s": 22.32, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Both stand: Engineering measures manufacturing lead time (confirmed order to despatch); Sales measures customer delivery lead time (purchase order to site delivery). They measure different fulfillment phases.\",\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Different metrics, different scope. No precedence rule in memory.\"\n}\n```\n", "obj": {"answer": "Both stand: Engineering measures manufacturing lead time (confirmed order to despatch); Sales measures customer delivery lead time (purchase order to site delivery). They measure different fulfillment phases.", "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Different metrics, different scope. No precedence rule in memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "HIST-05", "family": "HIST", "rep": 1, "latency_s": 16.32, "context_chars": 8932, "raw": "```json\n{\"answer\": \"No\", \"owner_team\": \"Finance\", \"source_id\": \"D-101\", \"policy_id\": null, \"unanswerable\": false, \"stale\": true, \"note\": \"D-101 valid on date. Superseded by D-102 on 2026-04-01.\"}\n```\n", "obj": {"answer": "No", "owner_team": "Finance", "source_id": "D-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "D-101 valid on date. Superseded by D-102 on 2026-04-01."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "CONF-02", "family": "CONF", "rep": 1, "latency_s": 23.12, "context_chars": 2694, "raw": "```json\n{\"answer\": null, \"owner_team\": null, \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"Sales and Support use different active_customer definitions; memory provides no resolution policy.\"}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Sales and Support use different active_customer definitions; memory provides no resolution policy."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "STALE-01", "family": "STALE", "rep": 0, "latency_s": 14.94, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"No. Legal, which owns this definition, changed it to five years effective 2026-06-01.\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"LEG-POL-04\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Support's seven-year version predates Legal's authoritative five-year policy.\"\n}\n```\n", "obj": {"answer": "No. Legal, which owns this definition, changed it to five years effective 2026-06-01.", "owner_team": "Legal", "source_id": "LEG-POL-04", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "Support's seven-year version predates Legal's authoritative five-year policy."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dimension", "item": "HIST-04", "family": "HIST", "rep": 1, "latency_s": 15.16, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"Version 2: an outage affecting any customer with no workaround, or any confirmed data loss\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-402 (authoritative, support-owned) applied from 2026-01-15 onwards.\"\n}\n```\n", "obj": {"answer": "Version 2: an outage affecting any customer with no workaround, or any confirmed data loss", "owner_team": "support", "source_id": "D-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "D-402 (authoritative, support-owned) applied from 2026-01-15 onwards."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "AUTH-02", "family": "AUTH", "rep": 0, "latency_s": 8.47, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Sales, specifically the VP Sales\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Stated in Who owns what section of shared wiki.\"\n}\n```\n", "obj": {"answer": "Sales, specifically the VP Sales", "owner_team": "Sales", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Stated in Who owns what section of shared wiki."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "ROUTE-03", "family": "ROUTE", "rep": 0, "latency_s": 17.78, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Ask the Legal team; the General Counsel is the accountable role for customer data retention policy.\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"LEG-POL-04\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal owns data_retention concept per governance catalogue\"\n}\n```\n", "obj": {"answer": "Ask the Legal team; the General Counsel is the accountable role for customer data retention policy.", "owner_team": "Legal", "source_id": "LEG-POL-04", "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal owns data_retention concept per governance catalogue"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "AUTH-05", "family": "AUTH", "rep": 1, "latency_s": 12.58, "context_chars": 11191, "raw": "```json\n{\"answer\": \"23 working days\", \"owner_team\": \"engineering\", \"source_id\": \"N-401\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Engineering's authoritative lead time figure from June 2026.\"}\n```\n", "obj": {"answer": "23 working days", "owner_team": "engineering", "source_id": "N-401", "policy_id": null, "unanswerable": false, "stale": false, "note": "Engineering's authoritative lead time figure from June 2026."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "HIST-02", "family": "HIST", "rep": 1, "latency_s": 26.55, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"customer records are kept for seven years after the contract ends\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"D-801\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-801 applied on 2026-03-01. Superseded by D-802 from 2026-06-01.\"\n}\n```\n", "obj": {"answer": "customer records are kept for seven years after the contract ends", "owner_team": "Legal", "source_id": "D-801", "policy_id": null, "unanswerable": false, "stale": true, "note": "D-801 applied on 2026-03-01. Superseded by D-802 from 2026-06-01."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "GAP-04", "family": "GAP", "rep": 0, "latency_s": 10.63, "context_chars": 11191, "raw": "```json\n{\"answer\": null, \"owner_team\": null, \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"Net promoter score not found in organizational memory\"}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Net promoter score not found in organizational memory"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "bitemporal", "item": "STALE-01", "family": "STALE", "rep": 1, "latency_s": 49.28, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"No. The company position is five years per Legal policy D-802 (effective 2026-06-01), not seven.\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": \"D-802\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Support wiki position contradicted by Legal D-802 same day: 7 vs 5 years retention.\"\n}\n```\n", "obj": {"answer": "No. The company position is five years per Legal policy D-802 (effective 2026-06-01), not seven.", "owner_team": "Legal", "source_id": "D-802", "policy_id": "D-802", "unanswerable": false, "stale": true, "note": "Support wiki position contradicted by Legal D-802 same day: 7 vs 5 years retention."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "graph_catalog", "item": "CONF-01", "family": "CONF", "rep": 1, "latency_s": 18.53, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"9.8m EUR (Finance, N-102) stands; Finance owns ARR and CP-ORG grants ownership prevalence\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"N-102\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales figure N-101 expired 2026-04-30; Finance N-102 remains valid\"\n}\n```\n", "obj": {"answer": "9.8m EUR (Finance, N-102) stands; Finance owns ARR and CP-ORG grants ownership prevalence", "owner_team": "Finance", "source_id": "N-102", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Sales figure N-101 expired 2026-04-30; Finance N-102 remains valid"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "CONF-04", "family": "CONF", "rep": 0, "latency_s": 17.58, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Five years after the contract ends\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"LEG-POL-04\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal owns definition; Version 2 from 2026-06-01 prevails\"\n}\n```\n", "obj": {"answer": "Five years after the contract ends", "owner_team": "Legal", "source_id": "LEG-POL-04", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Legal owns definition; Version 2 from 2026-06-01 prevails"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "CONF-01", "family": "CONF", "rep": 1, "latency_s": 41.45, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"9.8m EUR (Finance) per D-102 definition excluding one-off services and notice period customers\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"N-102\",\n  \"policy_id\": \"D-102\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales figure N-101 stale as of 2026-04-30; Finance figure remains valid\"\n}\n```\n", "obj": {"answer": "9.8m EUR (Finance) per D-102 definition excluding one-off services and notice period customers", "owner_team": "Finance", "source_id": "N-102", "policy_id": "D-102", "unanswerable": false, "stale": false, "note": "Sales figure N-101 stale as of 2026-04-30; Finance figure remains valid"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "STALE-04", "family": "STALE", "rep": 0, "latency_s": 14.93, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"a lead with a named budget holder and a stated timeframe\",\n  \"owner_team\": \"sales\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Definition from 2025-10-01 remains current with no superseding updates\"\n}\n```\n", "obj": {"answer": "a lead with a named budget holder and a stated timeframe", "owner_team": "sales", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Definition from 2025-10-01 remains current with no superseding updates"}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dimension", "item": "CONF-04", "family": "CONF", "rep": 0, "latency_s": 17.63, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"five years after the contract ends\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal's authoritative definition prevails; Support's record retained as their view.\"\n}\n```\n", "obj": {"answer": "five years after the contract ends", "owner_team": "legal", "source_id": "D-802", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Legal's authoritative definition prevails; Support's record retained as their view."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "CONF-03", "family": "CONF", "rep": 0, "latency_s": 19.69, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Engineering's 23 working days (N-401) stands; Engineering owns the lead_time concept per governance.\",\n  \"owner_team\": \"Engineering\",\n  \"source_id\": \"N-401\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Both records valid since 2026-06-01. Engineering owns lead_time; their record prevails.\"\n}\n```\n", "obj": {"answer": "Engineering's 23 working days (N-401) stands; Engineering owns the lead_time concept per governance.", "owner_team": "Engineering", "source_id": "N-401", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Both records valid since 2026-06-01. Engineering owns lead_time; their record prevails."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "GAP-04", "family": "GAP", "rep": 1, "latency_s": 11.44, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Net promoter score not found in organizational memory.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Net promoter score not found in organizational memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "graph_catalog", "item": "HIST-01", "family": "HIST", "rep": 1, "latency_s": 16.99, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"committed subscription value recognised over the next twelve months, excluding one-off services\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-101 was valid; competing D-103 from Sales overridden by Finance ownership rule.\"\n}\n```\n", "obj": {"answer": "committed subscription value recognised over the next twelve months, excluding one-off services", "owner_team": "Finance", "source_id": "D-101", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "D-101 was valid; competing D-103 from Sales overridden by Finance ownership rule."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "STALE-04", "family": "STALE", "rep": 0, "latency_s": 11.82, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Yes, Sales' definition of a qualified lead is still in force. Version 1, effective from 2025-10-01, remains current.\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"SAL-POL-06\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Definition: a lead with a named budget holder and a stated timeframe.\"\n}\n```\n", "obj": {"answer": "Yes, Sales' definition of a qualified lead is still in force. Version 1, effective from 2025-10-01, remains current.", "owner_team": "Sales", "source_id": "SAL-POL-06", "policy_id": null, "unanswerable": false, "stale": false, "note": "Definition: a lead with a named budget holder and a stated timeframe."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "vector_memory", "item": "ROUTE-03", "family": "ROUTE", "rep": 1, "latency_s": 17.1, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Legal team\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"2026-06-01 legal data_retention\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal owns data retention policy. Current period: 5 years post-contract.\"\n}\n```\n", "obj": {"answer": "Legal team", "owner_team": "legal", "source_id": "2026-06-01 legal data_retention", "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal owns data retention policy. Current period: 5 years post-contract."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "PROV-02", "family": "PROV", "rep": 0, "latency_s": 20.7, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Sales team, in document SAL-DECK-Q1\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"N-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Figure valid through 2026-04-30; superseded by Finance's 9.8m EUR\"\n}\n```\n", "obj": {"answer": "Sales team, in document SAL-DECK-Q1", "owner_team": "sales", "source_id": "N-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "Figure valid through 2026-04-30; superseded by Finance's 9.8m EUR"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "GAP-04", "family": "GAP", "rep": 1, "latency_s": 11.21, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"No record of net promoter score definition in shared memory.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "No record of net promoter score definition in shared memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dimension", "item": "STALE-01", "family": "STALE", "rep": 0, "latency_s": 20.67, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"No. Legal updated the company position on 2026-06-01 to five years (D-802). The Support wiki entry (D-803) is marked non-authoritative and has been superseded.\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Support wiki contradicts authoritative legal policy from June 2026\"\n}\n```\n", "obj": {"answer": "No. Legal updated the company position on 2026-06-01 to five years (D-802). The Support wiki entry (D-803) is marked non-authoritative and has been superseded.", "owner_team": "legal", "source_id": "D-802", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "Support wiki contradicts authoritative legal policy from June 2026"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "graph_catalog", "item": "AUTH-02", "family": "AUTH", "rep": 0, "latency_s": 30.15, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Sales\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"SAL-POL-02\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Governance catalogue lists Sales as owner, VP Sales accountable\"\n}\n```\n", "obj": {"answer": "Sales", "owner_team": "Sales", "source_id": "SAL-POL-02", "policy_id": null, "unanswerable": false, "stale": false, "note": "Governance catalogue lists Sales as owner, VP Sales accountable"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "PROV-01", "family": "PROV", "rep": 1, "latency_s": 21.38, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": \"finance\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Current definition held by finance since 2026-04-01, excludes one-off services and notice period customers.\"\n}\n```\n", "obj": {"answer": null, "owner_team": "finance", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Current definition held by finance since 2026-04-01, excludes one-off services and notice period customers."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "HIST-01", "family": "HIST", "rep": 0, "latency_s": 15.75, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"1\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-101 v1 valid 2025-01-01 to 2026-03-31. Superseded by D-102 v2 from 2026-04-01.\"\n}\n```\n", "obj": {"answer": "1", "owner_team": "finance", "source_id": "D-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "D-101 v1 valid 2025-01-01 to 2026-03-31. Superseded by D-102 v2 from 2026-04-01."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "ROUTE-05", "family": "ROUTE", "rep": 1, "latency_s": 28.02, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Sales and Engineering both provided definitions; Sales (D-602) is more recent\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"D-602\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Two valid definitions coexist. Sales' definition (2025-07-01) supersedes Engineering's.\"\n}\n```\n", "obj": {"answer": "Sales and Engineering both provided definitions; Sales (D-602) is more recent", "owner_team": "Sales", "source_id": "D-602", "policy_id": null, "unanswerable": false, "stale": false, "note": "Two valid definitions coexist. Sales' definition (2025-07-01) supersedes Engineering's."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "GAP-02", "family": "GAP", "rep": 1, "latency_s": 8.24, "context_chars": 6474, "raw": "```json\n{\"answer\": null, \"owner_team\": null, \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"No total headcount figure recorded in memory\"}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "No total headcount figure recorded in memory"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dimension", "item": "GAP-03", "family": "GAP", "rep": 1, "latency_s": 11.45, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"No definition of gross margin found in organizational memory.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "No definition of gross margin found in organizational memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "wiki", "item": "HIST-02", "family": "HIST", "rep": 1, "latency_s": 12.3, "context_chars": 6618, "raw": "```json\n{\"answer\": \"customer records are kept for seven years after the contract ends\", \"owner_team\": \"Legal\", \"source_id\": \"LEG-POL-01\", \"policy_id\": null, \"unanswerable\": false, \"stale\": true, \"note\": \"Version 1 superseded by Version 2 on 2026-06-01\"}\n```\n", "obj": {"answer": "customer records are kept for seven years after the contract ends", "owner_team": "Legal", "source_id": "LEG-POL-01", "policy_id": null, "unanswerable": false, "stale": true, "note": "Version 1 superseded by Version 2 on 2026-06-01"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "CONF-06", "family": "CONF", "rep": 0, "latency_s": 34.29, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Finance's definition applies (more recent): any person invoicing the company who is not on payroll\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-502\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance definition later than People's (2025-08-01 vs 2024-04-01); both currently valid\"\n}\n```\n", "obj": {"answer": "Finance's definition applies (more recent): any person invoicing the company who is not on payroll", "owner_team": "Finance", "source_id": "D-502", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance definition later than People's (2025-08-01 vs 2024-04-01); both currently valid"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 1, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "CONF-01", "family": "CONF", "rep": 1, "latency_s": 23.59, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Finance's 9.8m EUR figure (N-102) stands. It applies the current rule D-102: committed subscription value for next 12 months, excluding one-off services and customers in notice period, valid from 2026-04-01 onwards.\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"N-102\",\n  \"policy_id\": \"D-102\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales figure (N-101: 12.4m) expired 2026-04-30. Finance figure current.\"\n}\n```\n", "obj": {"answer": "Finance's 9.8m EUR figure (N-102) stands. It applies the current rule D-102: committed subscription value for next 12 months, excluding one-off services and customers in notice period, valid from 2026-04-01 onwards.", "owner_team": "finance", "source_id": "N-102", "policy_id": "D-102", "unanswerable": false, "stale": false, "note": "Sales figure (N-101: 12.4m) expired 2026-04-30. Finance figure current."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "CONF-02", "family": "CONF", "rep": 1, "latency_s": 56.66, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": \"N-201, N-202\",\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Two definitions valid simultaneously; no precedence rule provided\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": "N-201, N-202", "policy_id": null, "unanswerable": true, "stale": false, "note": "Two definitions valid simultaneously; no precedence rule provided"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "AUTH-01", "family": "AUTH", "rep": 0, "latency_s": 14.82, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"9.8m EUR\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"N-102\",\n  \"policy_id\": \"D-102\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance figure valid from 2026-01-01 ongoing; Sales figure expired 2026-04-30.\"\n}\n```\n", "obj": {"answer": "9.8m EUR", "owner_team": "Finance", "source_id": "N-102", "policy_id": "D-102", "unanswerable": false, "stale": false, "note": "Finance figure valid from 2026-01-01 ongoing; Sales figure expired 2026-04-30."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "PROV-03", "family": "PROV", "rep": 0, "latency_s": 48.09, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"LEG-POL-04\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-802 (LEG-POL-04) is authoritative from legal. D-803 from support conflicts but is non-authoritative.\"\n}\n```\n", "obj": {"answer": "LEG-POL-04", "owner_team": "legal", "source_id": "D-802", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "D-802 (LEG-POL-04) is authoritative from legal. D-803 from support conflicts but is non-authoritative."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "AUTH-06", "family": "AUTH", "rep": 0, "latency_s": 18.94, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"The people team owns the definition of a contractor\",\n  \"owner_team\": \"people\",\n  \"source_id\": \"D-501\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Head of People owns contractor definition; designated in concept_ownership\"\n}\n```\n", "obj": {"answer": "The people team owns the definition of a contractor", "owner_team": "people", "source_id": "D-501", "policy_id": null, "unanswerable": false, "stale": false, "note": "Head of People owns contractor definition; designated in concept_ownership"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "AUTH-02", "family": "AUTH", "rep": 1, "latency_s": 26.65, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Support owns the most recent definition; Sales provided an earlier version\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"D-202\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Two valid conflicting definitions exist; chose most recent\"\n}\n```\n", "obj": {"answer": "Support owns the most recent definition; Sales provided an earlier version", "owner_team": "Support", "source_id": "D-202", "policy_id": null, "unanswerable": false, "stale": false, "note": "Two valid conflicting definitions exist; chose most recent"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "GAP-01", "family": "GAP", "rep": 1, "latency_s": 12.0, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": \"finance\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory contains only 2026 ARR figures, no 2027 target\"\n}\n```\n", "obj": {"answer": null, "owner_team": "finance", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains only 2026 ARR figures, no 2027 target"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "wiki", "item": "PROV-02", "family": "PROV", "rep": 0, "latency_s": 21.25, "context_chars": 6618, "raw": "```json\n{\"answer\": \"Sales\", \"owner_team\": \"Finance\", \"source_id\": \"SAL-DECK-Q1\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Sales reported 12.4m EUR for period 2026-01-01 to 2026-04-30\"}\n```\n", "obj": {"answer": "Sales", "owner_team": "Finance", "source_id": "SAL-DECK-Q1", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales reported 12.4m EUR for period 2026-01-01 to 2026-04-30"}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "AUTH-01", "family": "AUTH", "rep": 1, "latency_s": 18.31, "context_chars": 8932, "raw": "```json\n{\"answer\": \"Finance: 9.8m EUR\", \"owner_team\": \"Finance\", \"source_id\": \"N-102\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Finance figure valid 2026-01-01 onwards; Sales figure expired 2026-04-30\"}\n```\n", "obj": {"answer": "Finance: 9.8m EUR", "owner_team": "Finance", "source_id": "N-102", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance figure valid 2026-01-01 onwards; Sales figure expired 2026-04-30"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "ROUTE-02", "family": "ROUTE", "rep": 1, "latency_s": 24.36, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Support team, as they own the current definition via their policy (D-402). However, Engineering also defines severity 1 (D-403), suggesting coordination may be needed.\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": \"D-403\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Conflict: Support policy (D-402) and Engineering wiki (D-403) have differing severity 1 definitions.\"\n}\n```\n", "obj": {"answer": "Support team, as they own the current definition via their policy (D-402). However, Engineering also defines severity 1 (D-403), suggesting coordination may be needed.", "owner_team": "support", "source_id": "D-402", "policy_id": "D-403", "unanswerable": false, "stale": false, "note": "Conflict: Support policy (D-402) and Engineering wiki (D-403) have differing severity 1 definitions."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "GAP-01", "family": "GAP", "rep": 0, "latency_s": 11.68, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": \"finance\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory contains 2026 ARR actuals but no 2027 target.\"\n}\n```\n", "obj": {"answer": null, "owner_team": "finance", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains 2026 ARR actuals but no 2027 target."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "STALE-01", "family": "STALE", "rep": 1, "latency_s": 26.53, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"No. The company position changed to five years retention as of 2026-06-01 per Legal policy D-802.\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": \"D-802\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Support wiki entry D-803 has not been updated to reflect the policy change from Legal D-802\"\n}\n```\n", "obj": {"answer": "No. The company position changed to five years retention as of 2026-06-01 per Legal policy D-802.", "owner_team": "legal", "source_id": "D-802", "policy_id": "D-802", "unanswerable": false, "stale": true, "note": "Support wiki entry D-803 has not been updated to reflect the policy change from Legal D-802"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dim_no_gov", "item": "AUTH-02", "family": "AUTH", "rep": 1, "latency_s": 46.84, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Sales (contract-based) and Support (activity-based); conflicting active definitions\",\n  \"owner_team\": \"Sales and Support\",\n  \"source_id\": \"D-201, D-202\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Two active definitions with conflicting criteria both currently in effect\"\n}\n```\n", "obj": {"answer": "Sales (contract-based) and Support (activity-based); conflicting active definitions", "owner_team": "Sales and Support", "source_id": "D-201, D-202", "policy_id": null, "unanswerable": false, "stale": false, "note": "Two active definitions with conflicting criteria both currently in effect"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "PROV-01", "family": "PROV", "rep": 0, "latency_s": 23.18, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"FIN-POL-07\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-102\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance owns ARR; D-102 (valid 2026-04-01, open) prevails over D-103.\"\n}\n```\n", "obj": {"answer": "FIN-POL-07", "owner_team": "Finance", "source_id": "D-102", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Finance owns ARR; D-102 (valid 2026-04-01, open) prevails over D-103."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "AUTH-04", "family": "AUTH", "rep": 1, "latency_s": 16.5, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"General Counsel\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"LEG-POL-04\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Accountable role within Legal team for data_retention\"\n}\n```\n", "obj": {"answer": "General Counsel", "owner_team": "Legal", "source_id": "LEG-POL-04", "policy_id": null, "unanswerable": false, "stale": false, "note": "Accountable role within Legal team for data_retention"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "AUTH-05", "family": "AUTH", "rep": 1, "latency_s": 25.58, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"41 calendar days\",\n  \"owner_team\": \"sales\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales lead time from PO to delivery on site; engineering also measures 23 working days from order to despatch.\"\n}\n```\n", "obj": {"answer": "41 calendar days", "owner_team": "sales", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales lead time from PO to delivery on site; engineering also measures 23 working days from order to despatch."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 1, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "AUTH-04", "family": "AUTH", "rep": 0, "latency_s": 20.23, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Legal\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal recorded current v2 policy. Support also records conflicting definition.\"\n}\n```\n", "obj": {"answer": "Legal", "owner_team": "legal", "source_id": "D-802", "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal recorded current v2 policy. Support also records conflicting definition."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "HIST-01", "family": "HIST", "rep": 0, "latency_s": 55.44, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"committed subscription value recognised over the next twelve months, excluding one-off services\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Finance's D-101 in force; Sales' D-103 also concurrent; D-101 later superseded by D-102.\"\n}\n```\n", "obj": {"answer": "committed subscription value recognised over the next twelve months, excluding one-off services", "owner_team": "finance", "source_id": "D-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "Finance's D-101 in force; Sales' D-103 also concurrent; D-101 later superseded by D-102."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "HIST-02", "family": "HIST", "rep": 1, "latency_s": 11.86, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"customer records are kept for seven years after the contract ends\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"D-801\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-801 was valid 2024-01-01 to 2026-05-31. Later superseded by D-802.\"\n}\n```\n", "obj": {"answer": "customer records are kept for seven years after the contract ends", "owner_team": "Legal", "source_id": "D-801", "policy_id": null, "unanswerable": false, "stale": false, "note": "D-801 was valid 2024-01-01 to 2026-05-31. Later superseded by D-802."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "HIST-05", "family": "HIST", "rep": 1, "latency_s": 17.33, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"No\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-101 valid 2025-12-31 excluded only one-off services; D-102 added notice period exclusion from 2026-04-01\"\n}\n```\n", "obj": {"answer": "No", "owner_team": "finance", "source_id": "D-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "D-101 valid 2025-12-31 excluded only one-off services; D-102 added notice period exclusion from 2026-04-01"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "ROUTE-05", "family": "ROUTE", "rep": 1, "latency_s": 30.46, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Sales team\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"D-602\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales defines order lead time as calendar days from purchase order to delivery. Engineering has alternative definition.\"\n}\n```\n", "obj": {"answer": "Sales team", "owner_team": "sales", "source_id": "D-602", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales defines order lead time as calendar days from purchase order to delivery. Engineering has alternative definition."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "ROUTE-01", "family": "ROUTE", "rep": 0, "latency_s": 27.39, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"Head of Finance\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-102\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance owns ARR concept; Head of Finance must approve\"\n}\n```\n", "obj": {"answer": "Head of Finance", "owner_team": "finance", "source_id": "D-102", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance owns ARR concept; Head of Finance must approve"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "CONF-05", "family": "CONF", "rep": 0, "latency_s": 12.58, "context_chars": 10621, "raw": "```json\n{\"answer\": \"an outage affecting any customer with no workaround, or any confirmed data loss\", \"owner_team\": \"Support\", \"source_id\": \"D-402\", \"policy_id\": \"CP-ORG\", \"unanswerable\": false, \"stale\": false, \"note\": \"Support owns concept; their current definition prevails over Engineering's.\"}\n```\n", "obj": {"answer": "an outage affecting any customer with no workaround, or any confirmed data loss", "owner_team": "Support", "source_id": "D-402", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Support owns concept; their current definition prevails over Engineering's."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "AUTH-04", "family": "AUTH", "rep": 0, "latency_s": 34.95, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Legal\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal policy current; Support has conflicting definition\"\n}\n```\n", "obj": {"answer": "Legal", "owner_team": "Legal", "source_id": "D-802", "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal policy current; Support has conflicting definition"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "AUTH-03", "family": "AUTH", "rep": 0, "latency_s": 44.39, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"318\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"SAL-REP-JUL\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Authoritative count from sales team, recorded July 2026\"\n}\n```\n", "obj": {"answer": "318", "owner_team": "sales", "source_id": "SAL-REP-JUL", "policy_id": null, "unanswerable": false, "stale": false, "note": "Authoritative count from sales team, recorded July 2026"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "HIST-05", "family": "HIST", "rep": 0, "latency_s": 16.14, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"No\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-101 was active then; excluded one-off services only, not notice period customers.\"\n}\n```\n", "obj": {"answer": "No", "owner_team": "finance", "source_id": "D-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "D-101 was active then; excluded one-off services only, not notice period customers."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "AUTH-06", "family": "AUTH", "rep": 1, "latency_s": 40.45, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Finance\",\n  \"owner_team\": \"finance\",\n  \"source_id\": null,\n  \"policy_id\": \"most recent definition is authoritative\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance redefined in 2025, superseding People's 2024 definition\"\n}\n```\n", "obj": {"answer": "Finance", "owner_team": "finance", "source_id": null, "policy_id": "most recent definition is authoritative", "unanswerable": false, "stale": false, "note": "Finance redefined in 2025, superseding People's 2024 definition"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "HIST-01", "family": "HIST", "rep": 1, "latency_s": 49.6, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"committed subscription value recognised over the next twelve months, excluding one-off services\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"2025-01-01\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance's definition; sales defined ARR differently on 2025-06-01\"\n}\n```\n", "obj": {"answer": "committed subscription value recognised over the next twelve months, excluding one-off services", "owner_team": "finance", "source_id": "2025-01-01", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance's definition; sales defined ARR differently on 2025-06-01"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "CONF-06", "family": "CONF", "rep": 1, "latency_s": 22.88, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Finance's definition applies: any person invoicing the company who is not on payroll\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-502\",\n  \"policy_id\": \"D-502\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance definition (Aug 2025) more recent than People's (Apr 2024); both valid\"\n}\n```\n", "obj": {"answer": "Finance's definition applies: any person invoicing the company who is not on payroll", "owner_team": "finance", "source_id": "D-502", "policy_id": "D-502", "unanswerable": false, "stale": false, "note": "Finance definition (Aug 2025) more recent than People's (Apr 2024); both valid"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 1, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "GAP-01", "family": "GAP", "rep": 1, "latency_s": 11.89, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"No ARR target for 2027 found in company records.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "No ARR target for 2027 found in company records."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "wiki", "item": "PROV-03", "family": "PROV", "rep": 1, "latency_s": 13.46, "context_chars": 6618, "raw": "```json\n{\"answer\": \"LEG-POL-04\", \"owner_team\": \"Legal\", \"source_id\": \"LEG-POL-04\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"LEG-POL-04 version 2 in force from 2026-06-01, still current\"}\n```\n", "obj": {"answer": "LEG-POL-04", "owner_team": "Legal", "source_id": "LEG-POL-04", "policy_id": null, "unanswerable": false, "stale": false, "note": "LEG-POL-04 version 2 in force from 2026-06-01, still current"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "CONF-05", "family": "CONF", "rep": 0, "latency_s": 60.59, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": \"D-402, D-403\",\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Support (D-402) and Engineering (D-403) have conflicting active definitions\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": "D-402, D-403", "policy_id": null, "unanswerable": true, "stale": false, "note": "Support (D-402) and Engineering (D-403) have conflicting active definitions"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "CONF-02", "family": "CONF", "rep": 0, "latency_s": 23.15, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": \"N-201, N-202\",\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Sales and Support use different active_customer definitions; both figures are valid but measure different populations.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": "N-201, N-202", "policy_id": null, "unanswerable": true, "stale": false, "note": "Sales and Support use different active_customer definitions; both figures are valid but measure different populations."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "CONF-05", "family": "CONF", "rep": 1, "latency_s": 12.35, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"an outage affecting any customer with no workaround, or any confirmed data loss\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support owns severity 1 incident. Their current definition (D-402, valid 2026-01-15+) prevails over Engineering's conflicting definition.\"\n}\n```\n", "obj": {"answer": "an outage affecting any customer with no workaround, or any confirmed data loss", "owner_team": "Support", "source_id": "D-402", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Support owns severity 1 incident. Their current definition (D-402, valid 2026-01-15+) prevails over Engineering's conflicting definition."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "STALE-02", "family": "STALE", "rep": 1, "latency_s": 14.43, "context_chars": 11191, "raw": "```json\n{\"answer\": \"No. Engineering's 2025-11-01 definition (D-403) is non-authoritative. Support owns this concept and issued an authoritative definition (D-402) on 2026-01-15.\", \"owner_team\": \"support\", \"source_id\": \"D-402\", \"policy_id\": \"CP-ORG\", \"unanswerable\": false, \"stale\": true, \"note\": \"Engineering's definition is non-authoritative; Support's later authoritative definition prevails.\"}\n```\n", "obj": {"answer": "No. Engineering's 2025-11-01 definition (D-403) is non-authoritative. Support owns this concept and issued an authoritative definition (D-402) on 2026-01-15.", "owner_team": "support", "source_id": "D-402", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "Engineering's definition is non-authoritative; Support's later authoritative definition prevails."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dimension", "item": "AUTH-05", "family": "AUTH", "rep": 0, "latency_s": 14.97, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"23 working days\",\n  \"owner_team\": \"engineering\",\n  \"source_id\": \"N-401\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"N-401 is authoritative and valid from 2026-06-01 onward.\"\n}\n```\n", "obj": {"answer": "23 working days", "owner_team": "engineering", "source_id": "N-401", "policy_id": null, "unanswerable": false, "stale": false, "note": "N-401 is authoritative and valid from 2026-06-01 onward."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "GAP-02", "family": "GAP", "rep": 0, "latency_s": 9.59, "context_chars": 6474, "raw": "```json\n{\"answer\": null, \"owner_team\": \"people\", \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"Memory contains headcount plan but not current total headcount.\"}\n```\n", "obj": {"answer": null, "owner_team": "people", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains headcount plan but not current total headcount."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "CONF-06", "family": "CONF", "rep": 0, "latency_s": 18.06, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"any person invoicing the company who is not on payroll\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-502\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance definition D-502 (2025-08-01) more recent than People's D-501 (2024-04-01).\"\n}\n```\n", "obj": {"answer": "any person invoicing the company who is not on payroll", "owner_team": "finance", "source_id": "D-502", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance definition D-502 (2025-08-01) more recent than People's D-501 (2024-04-01)."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 1, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "GAP-01", "family": "GAP", "rep": 0, "latency_s": 11.64, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": \"finance\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory contains only 2026 ARR figures, no 2027 target recorded.\"\n}\n```\n", "obj": {"answer": null, "owner_team": "finance", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains only 2026 ARR figures, no 2027 target recorded."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "ROUTE-03", "family": "ROUTE", "rep": 1, "latency_s": 15.81, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Legal team\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal team recorded the current data retention policy definition.\"\n}\n```\n", "obj": {"answer": "Legal team", "owner_team": "legal", "source_id": "D-802", "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal team recorded the current data retention policy definition."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "AUTH-03", "family": "AUTH", "rep": 1, "latency_s": 11.03, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"318\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"N-201\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales-recorded authoritative figure; support's 402 count deferred.\"\n}\n```\n", "obj": {"answer": "318", "owner_team": "sales", "source_id": "N-201", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales-recorded authoritative figure; support's 402 count deferred."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "AUTH-01", "family": "AUTH", "rep": 0, "latency_s": 16.58, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"9.8m EUR\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"FIN-STMT-Q1\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance owns ARR definition; their current report prevails over Sales.\"\n}\n```\n", "obj": {"answer": "9.8m EUR", "owner_team": "Finance", "source_id": "FIN-STMT-Q1", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Finance owns ARR definition; their current report prevails over Sales."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "CONF-06", "family": "CONF", "rep": 1, "latency_s": 11.76, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"a person engaged under a services agreement with no employment relationship\",\n  \"owner_team\": \"People\",\n  \"source_id\": \"D-501\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"People owns contractor; their definition prevails per CP-ORG.\"\n}\n```\n", "obj": {"answer": "a person engaged under a services agreement with no employment relationship", "owner_team": "People", "source_id": "D-501", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "People owns contractor; their definition prevails per CP-ORG."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "HIST-03", "family": "HIST", "rep": 0, "latency_s": 22.74, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"an outage affecting more than one customer with no workaround\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"D-401\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Support-owned definition prevailed over Engineering's competing definition on that date.\"\n}\n```\n", "obj": {"answer": "an outage affecting more than one customer with no workaround", "owner_team": "Support", "source_id": "D-401", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "Support-owned definition prevailed over Engineering's competing definition on that date."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "STALE-03", "family": "STALE", "rep": 1, "latency_s": 12.55, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"No. Version 1 ended 2026-03-31; version 2 in force from 2026-04-01.\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"FIN-POL-04\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Version 1 superseded by version 2 from 2026-04-01.\"\n}\n```\n", "obj": {"answer": "No. Version 1 ended 2026-03-31; version 2 in force from 2026-04-01.", "owner_team": "Finance", "source_id": "FIN-POL-04", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "Version 1 superseded by version 2 from 2026-04-01."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "wiki", "item": "CONF-03", "family": "CONF", "rep": 1, "latency_s": 14.12, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Engineering's order lead time of 23 working days stands. Engineering owns the definition of order lead time.\",\n  \"owner_team\": \"Engineering\",\n  \"source_id\": \"ENG-REP-JUN\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Engineering owns the order lead time definition; their report prevails per CP-ORG.\"\n}\n```\n", "obj": {"answer": "Engineering's order lead time of 23 working days stands. Engineering owns the definition of order lead time.", "owner_team": "Engineering", "source_id": "ENG-REP-JUN", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Engineering owns the order lead time definition; their report prevails per CP-ORG."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "PROV-03", "family": "PROV", "rep": 0, "latency_s": 42.09, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"LEG-POL-04\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": \"D-802\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Two rules in force: Legal 5yr, Support 7yr. Legal preferred.\"\n}\n```\n", "obj": {"answer": "LEG-POL-04", "owner_team": "Legal", "source_id": "D-802", "policy_id": "D-802", "unanswerable": false, "stale": false, "note": "Two rules in force: Legal 5yr, Support 7yr. Legal preferred."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "ROUTE-02", "family": "ROUTE", "rep": 0, "latency_s": 14.8, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"The Support Lead. Support owns the definition of severity 1 incidents.\",\n  \"owner_team\": \"Support\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support owns severity 1 incident definition per org ownership rules.\"\n}\n```\n", "obj": {"answer": "The Support Lead. Support owns the definition of severity 1 incidents.", "owner_team": "Support", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Support owns severity 1 incident definition per org ownership rules."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "ROUTE-05", "family": "ROUTE", "rep": 1, "latency_s": 35.46, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Engineering; working days from confirmed order to despatch from the works\",\n  \"owner_team\": \"engineering\",\n  \"source_id\": \"2025-05-01\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales records a different metric: calendar days to customer site delivery\"\n}\n```\n", "obj": {"answer": "Engineering; working days from confirmed order to despatch from the works", "owner_team": "engineering", "source_id": "2025-05-01", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales records a different metric: calendar days to customer site delivery"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "ROUTE-04", "family": "ROUTE", "rep": 0, "latency_s": 15.66, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Sales (VP Sales accountable role)\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"SAL-POL-06\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales owns qualified_lead concept per governance catalogue; VP Sales is accountable role\"\n}\n```\n", "obj": {"answer": "Sales (VP Sales accountable role)", "owner_team": "Sales", "source_id": "SAL-POL-06", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales owns qualified_lead concept per governance catalogue; VP Sales is accountable role"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "PROV-01", "family": "PROV", "rep": 1, "latency_s": 20.27, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"FIN-POL-07\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-102\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Conflicting sales definition (SAL-WIKI-11) also currently valid; finance definition chosen as authoritative.\"\n}\n```\n", "obj": {"answer": "FIN-POL-07", "owner_team": "finance", "source_id": "D-102", "policy_id": null, "unanswerable": false, "stale": false, "note": "Conflicting sales definition (SAL-WIKI-11) also currently valid; finance definition chosen as authoritative."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "ROUTE-01", "family": "ROUTE", "rep": 1, "latency_s": 15.3, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Both Finance and Sales teams must agree. They publish conflicting ARR definitions: Finance defines it as subscription value excluding one-off services and notice-period customers; Sales includes one-off services over signed contracts.\",\n  \"owner_team\": \"finance\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance and Sales report different ARR values (9.8m vs 12.4m EUR)\"\n}\n```\n", "obj": {"answer": "Both Finance and Sales teams must agree. They publish conflicting ARR definitions: Finance defines it as subscription value excluding one-off services and notice-period customers; Sales includes one-off services over signed contracts.", "owner_team": "finance", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance and Sales report different ARR values (9.8m vs 12.4m EUR)"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "AUTH-01", "family": "AUTH", "rep": 1, "latency_s": 51.29, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Finance: 9.8m EUR\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"FIN-STMT-Q1\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance owns definition; Finance figure prevails over Sales figure per CP-ORG.\"\n}\n```\n", "obj": {"answer": "Finance: 9.8m EUR", "owner_team": "Finance", "source_id": "FIN-STMT-Q1", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Finance owns definition; Finance figure prevails over Sales figure per CP-ORG."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "AUTH-04", "family": "AUTH", "rep": 1, "latency_s": 41.98, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Legal\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal owns D-802 policy; Support recorded conflicting D-803\"\n}\n```\n", "obj": {"answer": "Legal", "owner_team": "legal", "source_id": "D-802", "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal owns D-802 policy; Support recorded conflicting D-803"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "PROV-02", "family": "PROV", "rep": 1, "latency_s": 13.43, "context_chars": 2694, "raw": "```json\n{\"answer\": \"sales\", \"owner_team\": \"sales\", \"source_id\": null, \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Sales team reported 12.4m EUR on 2026-01-01\"}\n```\n", "obj": {"answer": "sales", "owner_team": "sales", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales team reported 12.4m EUR on 2026-01-01"}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "ROUTE-02", "family": "ROUTE", "rep": 0, "latency_s": 15.39, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Support (Support Lead is the accountable role)\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"SUP-POL-03\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support owns severity 1 incident concept per governance catalogue\"\n}\n```\n", "obj": {"answer": "Support (Support Lead is the accountable role)", "owner_team": "Support", "source_id": "SUP-POL-03", "policy_id": null, "unanswerable": false, "stale": false, "note": "Support owns severity 1 incident concept per governance catalogue"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "PROV-01", "family": "PROV", "rep": 1, "latency_s": 8.95, "context_chars": 6618, "raw": "```json\n{\"answer\": \"FIN-POL-07\", \"owner_team\": \"Finance\", \"source_id\": \"FIN-POL-07\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Version 2, in force from 2026-04-01\"}\n```\n", "obj": {"answer": "FIN-POL-07", "owner_team": "Finance", "source_id": "FIN-POL-07", "policy_id": null, "unanswerable": false, "stale": false, "note": "Version 2, in force from 2026-04-01"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "HIST-01", "family": "HIST", "rep": 1, "latency_s": 59.15, "context_chars": 6474, "raw": "```json\n{\"answer\": \"1\", \"owner_team\": \"finance\", \"source_id\": \"D-101\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Finance (D-101) and sales (D-103) both had version 1 in force on this date\"}\n```\n", "obj": {"answer": "1", "owner_team": "finance", "source_id": "D-101", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance (D-101) and sales (D-103) both had version 1 in force on this date"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "CONF-04", "family": "CONF", "rep": 1, "latency_s": 25.17, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Legal's policy (D-802): five years after contract ends. Support contradicts with seven years (D-803).\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Both effective 2026-06-01; Legal is authoritative; active contradiction.\"\n}\n```\n", "obj": {"answer": "Legal's policy (D-802): five years after contract ends. Support contradicts with seven years (D-803).", "owner_team": "Legal", "source_id": "D-802", "policy_id": null, "unanswerable": false, "stale": false, "note": "Both effective 2026-06-01; Legal is authoritative; active contradiction."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "PROV-02", "family": "PROV", "rep": 1, "latency_s": 23.59, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Sales\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"N-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Sales reported 12.4m EUR (SAL-DECK-Q1, valid 2026-01-01 to 2026-04-30). Now stale.\"\n}\n```\n", "obj": {"answer": "Sales", "owner_team": "Finance", "source_id": "N-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "Sales reported 12.4m EUR (SAL-DECK-Q1, valid 2026-01-01 to 2026-04-30). Now stale."}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "HIST-03", "family": "HIST", "rep": 0, "latency_s": 14.96, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Version 1\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"SUP-POL-01\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Version 1 was in force 2025-03-01 to 2026-01-14. Now superseded by Version 2.\"\n}\n```\n", "obj": {"answer": "Version 1", "owner_team": "Support", "source_id": "SUP-POL-01", "policy_id": null, "unanswerable": false, "stale": true, "note": "Version 1 was in force 2025-03-01 to 2026-01-14. Now superseded by Version 2."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "ROUTE-05", "family": "ROUTE", "rep": 0, "latency_s": 40.72, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Sales is accountable; defined as calendar days from customer's purchase order to delivery on site\",\n  \"owner_team\": \"sales\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales defined order lead time 2025-07-01; Engineering defines separate manufacturing lead time metric\"\n}\n```\n", "obj": {"answer": "Sales is accountable; defined as calendar days from customer's purchase order to delivery on site", "owner_team": "sales", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales defined order lead time 2025-07-01; Engineering defines separate manufacturing lead time metric"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "ROUTE-03", "family": "ROUTE", "rep": 1, "latency_s": 14.88, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"Ask the General Counsel (legal team)\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal team owns customer data retention policy as of 2026-06-01\"\n}\n```\n", "obj": {"answer": "Ask the General Counsel (legal team)", "owner_team": "legal", "source_id": "D-802", "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal team owns customer data retention policy as of 2026-06-01"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "AUTH-04", "family": "AUTH", "rep": 0, "latency_s": 17.37, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Legal team is accountable for customer data retention policy\",\n  \"owner_team\": \"legal\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal policy most recently updated 2026-06-01 to five years post-contract end\"\n}\n```\n", "obj": {"answer": "Legal team is accountable for customer data retention policy", "owner_team": "legal", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal policy most recently updated 2026-06-01 to five years post-contract end"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "CONF-01", "family": "CONF", "rep": 0, "latency_s": 15.76, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Finance's 9.8m EUR figure stands: ARR is committed subscription value over next twelve months, excluding one-off services and customers in notice period\",\n  \"owner_team\": \"finance\",\n  \"source_id\": null,\n  \"policy_id\": \"2026-04-01\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales definition (12.4m) includes one-off services; Finance definition (9.8m) excludes them\"\n}\n```\n", "obj": {"answer": "Finance's 9.8m EUR figure stands: ARR is committed subscription value over next twelve months, excluding one-off services and customers in notice period", "owner_team": "finance", "source_id": null, "policy_id": "2026-04-01", "unanswerable": false, "stale": false, "note": "Sales definition (12.4m) includes one-off services; Finance definition (9.8m) excludes them"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "PROV-01", "family": "PROV", "rep": 0, "latency_s": 23.43, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"FIN-POL-07\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-102\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-102 (finance, authoritative, 2026-04-01 forward) prevails over D-103 (sales, non-authoritative)\"\n}\n```\n", "obj": {"answer": "FIN-POL-07", "owner_team": "finance", "source_id": "D-102", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "D-102 (finance, authoritative, 2026-04-01 forward) prevails over D-103 (sales, non-authoritative)"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "PROV-01", "family": "PROV", "rep": 0, "latency_s": 12.86, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"FIN-POL-07\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"FIN-POL-07\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance's Version 2 ARR definition in force from 2026-04-01, still current.\"\n}\n```\n", "obj": {"answer": "FIN-POL-07", "owner_team": "Finance", "source_id": "FIN-POL-07", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance's Version 2 ARR definition in force from 2026-04-01, still current."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "STALE-04", "family": "STALE", "rep": 1, "latency_s": 9.68, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Yes, a lead with a named budget holder and a stated timeframe\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"D-701\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Definition valid from 2025-10-01 with no end date.\"\n}\n```\n", "obj": {"answer": "Yes, a lead with a named budget holder and a stated timeframe", "owner_team": "Sales", "source_id": "D-701", "policy_id": null, "unanswerable": false, "stale": false, "note": "Definition valid from 2025-10-01 with no end date."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dimension", "item": "HIST-03", "family": "HIST", "rep": 1, "latency_s": 17.08, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"1\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"D-401\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-401 was authoritative; D-403 conflicted but non-authoritative.\"\n}\n```\n", "obj": {"answer": "1", "owner_team": "support", "source_id": "D-401", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "D-401 was authoritative; D-403 conflicted but non-authoritative."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "CONF-05", "family": "CONF", "rep": 0, "latency_s": 15.3, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"an outage affecting any customer with no workaround, or any confirmed data loss\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support owns severity 1 incidents. D-402 (support, authoritative) supersedes D-403 (engineering, non-authoritative).\"\n}\n```\n", "obj": {"answer": "an outage affecting any customer with no workaround, or any confirmed data loss", "owner_team": "support", "source_id": "D-402", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Support owns severity 1 incidents. D-402 (support, authoritative) supersedes D-403 (engineering, non-authoritative)."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "HIST-03", "family": "HIST", "rep": 0, "latency_s": 92.29, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"an outage affecting more than one customer with no workaround\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"D-401\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-401 primary from support; D-403 engineering criterion also active\"\n}\n```\n", "obj": {"answer": "an outage affecting more than one customer with no workaround", "owner_team": "support", "source_id": "D-401", "policy_id": null, "unanswerable": false, "stale": true, "note": "D-401 primary from support; D-403 engineering criterion also active"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "HIST-03", "family": "HIST", "rep": 0, "latency_s": 30.87, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"1\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"D-401\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-401 authoritative on 2025-12-01; superseded by D-402 on 2026-01-15\"\n}\n```\n", "obj": {"answer": "1", "owner_team": "support", "source_id": "D-401", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "D-401 authoritative on 2025-12-01; superseded by D-402 on 2026-01-15"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "HIST-02", "family": "HIST", "rep": 0, "latency_s": 12.65, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"customer records are kept for seven years after the contract ends\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-801\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-801 valid through 2026-05-31. Superseded by D-802 from 2026-06-01.\"\n}\n```\n", "obj": {"answer": "customer records are kept for seven years after the contract ends", "owner_team": "legal", "source_id": "D-801", "policy_id": null, "unanswerable": false, "stale": true, "note": "D-801 valid through 2026-05-31. Superseded by D-802 from 2026-06-01."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "AUTH-06", "family": "AUTH", "rep": 1, "latency_s": 52.41, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Finance team (definition recorded 2025-08-01); people team also recorded definition (2024-04-01)\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-502\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Two teams recorded conflicting definitions; finance has more recent version\"\n}\n```\n", "obj": {"answer": "Finance team (definition recorded 2025-08-01); people team also recorded definition (2024-04-01)", "owner_team": "finance", "source_id": "D-502", "policy_id": null, "unanswerable": false, "stale": false, "note": "Two teams recorded conflicting definitions; finance has more recent version"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "AUTH-06", "family": "AUTH", "rep": 1, "latency_s": 14.67, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"people\",\n  \"owner_team\": \"people\",\n  \"source_id\": \"D-501\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"people team owns contractor concept; Head of People is owner role\"\n}\n```\n", "obj": {"answer": "people", "owner_team": "people", "source_id": "D-501", "policy_id": null, "unanswerable": false, "stale": false, "note": "people team owns contractor concept; Head of People is owner role"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "CONF-01", "family": "CONF", "rep": 0, "latency_s": 24.15, "context_chars": 10621, "raw": "```json\n{\"answer\": \"9.8m EUR (Finance report N-102)\", \"owner_team\": \"Finance\", \"source_id\": \"N-102\", \"policy_id\": \"CP-ORG\", \"unanswerable\": false, \"stale\": true, \"note\": \"Finance owns ARR; CP-ORG rule applies. Sales figure expired 2026-04-30.\"}\n```\n", "obj": {"answer": "9.8m EUR (Finance report N-102)", "owner_team": "Finance", "source_id": "N-102", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "Finance owns ARR; CP-ORG rule applies. Sales figure expired 2026-04-30."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "GAP-04", "family": "GAP", "rep": 0, "latency_s": 9.72, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Net promoter score not found in organization memory.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Net promoter score not found in organization memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "bitemporal", "item": "STALE-03", "family": "STALE", "rep": 1, "latency_s": 13.92, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"No. Version 1 (D-101) expired 2026-03-31 and is superseded by D-102 (valid from 2026-04-01).\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"ARR v1 definition expired; v3 (D-102) now governs.\"\n}\n```\n", "obj": {"answer": "No. Version 1 (D-101) expired 2026-03-31 and is superseded by D-102 (valid from 2026-04-01).", "owner_team": "Finance", "source_id": "D-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "ARR v1 definition expired; v3 (D-102) now governs."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "vector_memory", "item": "STALE-05", "family": "STALE", "rep": 1, "latency_s": 34.25, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Yes. Finance defined churn on 2025-01-01 as customers whose contract ended in the period and was not renewed within thirty days. No superseding definition appears in the memory.\",\n  \"owner_team\": \"finance\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Memory labels it 'churn', not explicitly 'logo churn'\"\n}\n```\n", "obj": {"answer": "Yes. Finance defined churn on 2025-01-01 as customers whose contract ended in the period and was not renewed within thirty days. No superseding definition appears in the memory.", "owner_team": "finance", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Memory labels it 'churn', not explicitly 'logo churn'"}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dim_no_gov", "item": "ROUTE-05", "family": "ROUTE", "rep": 0, "latency_s": 75.92, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Sales\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"D-602\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-602 defines customer order-to-delivery lead time\"\n}\n```\n", "obj": {"answer": "Sales", "owner_team": "sales", "source_id": "D-602", "policy_id": null, "unanswerable": false, "stale": false, "note": "D-602 defines customer order-to-delivery lead time"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "ROUTE-04", "family": "ROUTE", "rep": 1, "latency_s": 8.26, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"The VP Sales\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"SAL-POL-06\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"VP Sales owns the qualified lead definition per organizational records.\"\n}\n```\n", "obj": {"answer": "The VP Sales", "owner_team": "Sales", "source_id": "SAL-POL-06", "policy_id": null, "unanswerable": false, "stale": false, "note": "VP Sales owns the qualified lead definition per organizational records."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "CONF-06", "family": "CONF", "rep": 1, "latency_s": 24.61, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Finance definition from 2025-08-01 applies: 'any person invoicing the company who is not on payroll'. People's 2024 definition required a services agreement with no employment relationship—substantively different.\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"2025-08-01\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance definition is more recent; no explicit policy resolves conflict.\"\n}\n```\n", "obj": {"answer": "Finance definition from 2025-08-01 applies: 'any person invoicing the company who is not on payroll'. People's 2024 definition required a services agreement with no employment relationship—substantively different.", "owner_team": "finance", "source_id": "2025-08-01", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance definition is more recent; no explicit policy resolves conflict."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "GAP-03", "family": "GAP", "rep": 0, "latency_s": 8.5, "context_chars": 6618, "raw": "```json\n{\"answer\": null, \"owner_team\": null, \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"Gross margin definition not found in shared memory.\"}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Gross margin definition not found in shared memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dimension", "item": "HIST-04", "family": "HIST", "rep": 0, "latency_s": 14.86, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"Version 2\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-402 effective from 2026-01-15; D-401 expired 2026-01-14\"\n}\n```\n", "obj": {"answer": "Version 2", "owner_team": "support", "source_id": "D-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "D-402 effective from 2026-01-15; D-401 expired 2026-01-14"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "GAP-02", "family": "GAP", "rep": 0, "latency_s": 11.37, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": \"people\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory contains engineering headcount growth plan (S-105) but not total company headcount.\"\n}\n```\n", "obj": {"answer": null, "owner_team": "people", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains engineering headcount growth plan (S-105) but not total company headcount."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "vector_memory", "item": "ROUTE-04", "family": "ROUTE", "rep": 0, "latency_s": 10.46, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"The sales team must be asked to change the definition of a qualified lead.\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"2025-10-01\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales team owns the qualified_lead concept definition.\"\n}\n```\n", "obj": {"answer": "The sales team must be asked to change the definition of a qualified lead.", "owner_team": "sales", "source_id": "2025-10-01", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales team owns the qualified_lead concept definition."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "AUTH-05", "family": "AUTH", "rep": 0, "latency_s": 30.59, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"41 calendar days\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"N-402\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Customer purchase order to on-site delivery metric; engineering also reports 23 working days (N-401)\"\n}\n```\n", "obj": {"answer": "41 calendar days", "owner_team": "sales", "source_id": "N-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "Customer purchase order to on-site delivery metric; engineering also reports 23 working days (N-401)"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 1, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "STALE-05", "family": "STALE", "rep": 1, "latency_s": 10.06, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"No definition of 'logo churn' found in memory records.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "No definition of 'logo churn' found in memory records."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "wiki", "item": "AUTH-05", "family": "AUTH", "rep": 0, "latency_s": 16.88, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"23 working days\",\n  \"owner_team\": \"Engineering\",\n  \"source_id\": \"ENG-REP-JUN\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Engineering owns definition. Sales measured calendar days; Engineering's working days prevails.\"\n}\n```\n", "obj": {"answer": "23 working days", "owner_team": "Engineering", "source_id": "ENG-REP-JUN", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Engineering owns definition. Sales measured calendar days; Engineering's working days prevails."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "AUTH-05", "family": "AUTH", "rep": 1, "latency_s": 15.35, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"23 working days\",\n  \"owner_team\": \"Engineering\",\n  \"source_id\": \"ENG-REP-JUN\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Engineering owns the concept; both teams reported conflicting figures for 2026-06-01.\"\n}\n```\n", "obj": {"answer": "23 working days", "owner_team": "Engineering", "source_id": "ENG-REP-JUN", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Engineering owns the concept; both teams reported conflicting figures for 2026-06-01."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "AUTH-06", "family": "AUTH", "rep": 0, "latency_s": 26.86, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"people\",\n  \"owner_team\": \"people\",\n  \"source_id\": \"D-501\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance also recorded a different definition (D-502)\"\n}\n```\n", "obj": {"answer": "people", "owner_team": "people", "source_id": "D-501", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance also recorded a different definition (D-502)"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "GAP-04", "family": "GAP", "rep": 0, "latency_s": 11.54, "context_chars": 2694, "raw": "```json\n{\"answer\": null, \"owner_team\": null, \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"Net promoter score definition not present in memory.\"}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Net promoter score definition not present in memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dimension", "item": "GAP-03", "family": "GAP", "rep": 0, "latency_s": 12.19, "context_chars": 11191, "raw": "```json\n{\"answer\": null, \"owner_team\": null, \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"No record of gross margin definition found in organizational memory.\"}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "No record of gross margin definition found in organizational memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "graph_catalog", "item": "ROUTE-04", "family": "ROUTE", "rep": 1, "latency_s": 14.26, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Sales\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"D-701\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales owns qualified lead concept; VP Sales is accountable role.\"\n}\n```\n", "obj": {"answer": "Sales", "owner_team": "Sales", "source_id": "D-701", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales owns qualified lead concept; VP Sales is accountable role."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "ROUTE-01", "family": "ROUTE", "rep": 0, "latency_s": 32.98, "context_chars": 6474, "raw": "```json\n{\"answer\": \"Finance and Sales must both agree. Both teams have recorded different active definitions of ARR: Finance defines it in D-102 (committed subscription value, excluding one-off services and customers in notice period, from 2026-04-01 onward) while Sales defines it in D-103 (total value of contracts signed, including one-off services, from 2025-06-01 onward). Changing ARR calculation requires aligning these conflicting definitions.\", \"owner_team\": \"Finance and Sales\", \"source_id\"", "obj": {"answer": "Finance and Sales must both agree. Both teams have recorded different active definitions of ARR: Finance defines it in D-102 (committed subscription value, excluding one-off services and customers in notice period, from 2026-04-01 onward) while Sales defines it in D-103 (total value of contracts signed, including one-off services, from 2025-06-01 onward). Changing ARR calculation requires aligning these conflicting definitions.", "owner_team": "Finance and Sales", "source_id": "D-102, D-103", "policy_id": null, "unanswerable": false, "stale": false, "note": "Multiple active, conflicting ARR definitions from two teams."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "PROV-03", "family": "PROV", "rep": 1, "latency_s": 17.76, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"LEG-POL-04\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": \"D-802\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-802 (5 years) chosen over conflicting D-803 (7 years); both valid from 2026-06-01\"\n}\n```\n", "obj": {"answer": "LEG-POL-04", "owner_team": "Legal", "source_id": "D-802", "policy_id": "D-802", "unanswerable": false, "stale": false, "note": "D-802 (5 years) chosen over conflicting D-803 (7 years); both valid from 2026-06-01"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "ROUTE-03", "family": "ROUTE", "rep": 1, "latency_s": 25.37, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Legal team (General Counsel)\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"LEG-POL-04\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"General Counsel is the accountable role for data retention policy\"\n}\n```\n", "obj": {"answer": "Legal team (General Counsel)", "owner_team": "Legal", "source_id": "LEG-POL-04", "policy_id": null, "unanswerable": false, "stale": false, "note": "General Counsel is the accountable role for data retention policy"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "ROUTE-05", "family": "ROUTE", "rep": 0, "latency_s": 12.14, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Engineering Manager\",\n  \"owner_team\": \"Engineering\",\n  \"source_id\": \"ENG-POL-02\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Engineering Manager owns the order lead time definition, in force from 2025-05-01\"\n}\n```\n", "obj": {"answer": "Engineering Manager", "owner_team": "Engineering", "source_id": "ENG-POL-02", "policy_id": null, "unanswerable": false, "stale": false, "note": "Engineering Manager owns the order lead time definition, in force from 2025-05-01"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "ROUTE-05", "family": "ROUTE", "rep": 1, "latency_s": 22.0, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Engineering Manager\",\n  \"owner_team\": \"Engineering\",\n  \"source_id\": \"ENG-POL-02\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Governance catalogue establishes Engineering as owner with Engineering Manager accountable\"\n}\n```\n", "obj": {"answer": "Engineering Manager", "owner_team": "Engineering", "source_id": "ENG-POL-02", "policy_id": null, "unanswerable": false, "stale": false, "note": "Governance catalogue establishes Engineering as owner with Engineering Manager accountable"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "AUTH-03", "family": "AUTH", "rep": 1, "latency_s": 38.29, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"318\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"N-201\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support also reported 402. Sales uses contract-based definition; support uses activity-based.\"\n}\n```\n", "obj": {"answer": "318", "owner_team": "sales", "source_id": "N-201", "policy_id": null, "unanswerable": false, "stale": false, "note": "Support also reported 402. Sales uses contract-based definition; support uses activity-based."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "HIST-05", "family": "HIST", "rep": 0, "latency_s": 16.54, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"No\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-101 excluded one-off services only, not notice period customers. D-102 added that exclusion on 2026-04-01.\"\n}\n```\n", "obj": {"answer": "No", "owner_team": "Finance", "source_id": "D-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "D-101 excluded one-off services only, not notice period customers. D-102 added that exclusion on 2026-04-01."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "CONF-02", "family": "CONF", "rep": 1, "latency_s": 16.02, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"318\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"N-201\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales owns concept. Both figures valid but differ by definition used.\"\n}\n```\n", "obj": {"answer": "318", "owner_team": "Sales", "source_id": "N-201", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Sales owns concept. Both figures valid but differ by definition used."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "AUTH-05", "family": "AUTH", "rep": 0, "latency_s": 22.25, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"41 calendar days\",\n  \"owner_team\": \"sales\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales lead time: customer PO to delivery on site. Engineering also records 23 working days.\"\n}\n```\n", "obj": {"answer": "41 calendar days", "owner_team": "sales", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales lead time: customer PO to delivery on site. Engineering also records 23 working days."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 1, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "GAP-01", "family": "GAP", "rep": 0, "latency_s": 10.72, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": \"Finance\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory contains 2026 actuals only, no 2027 target recorded.\"\n}\n```\n", "obj": {"answer": null, "owner_team": "Finance", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains 2026 actuals only, no 2027 target recorded."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "wiki", "item": "ROUTE-05", "family": "ROUTE", "rep": 1, "latency_s": 11.28, "context_chars": 6618, "raw": "```json\n{\"answer\": \"Engineering Manager\", \"owner_team\": \"Engineering\", \"source_id\": null, \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"From governance structure: Engineering Manager owns definition\"}\n```\n", "obj": {"answer": "Engineering Manager", "owner_team": "Engineering", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "From governance structure: Engineering Manager owns definition"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "CONF-03", "family": "CONF", "rep": 0, "latency_s": 19.66, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"Engineering's record stands: 23 working days from confirmed order to despatch. Sales reports 41 calendar days but their record is non-authoritative.\",\n  \"owner_team\": \"engineering\",\n  \"source_id\": \"N-401\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Engineering Manager owns lead_time concept; authoritative record prevails.\"\n}\n```\n", "obj": {"answer": "Engineering's record stands: 23 working days from confirmed order to despatch. Sales reports 41 calendar days but their record is non-authoritative.", "owner_team": "engineering", "source_id": "N-401", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Engineering Manager owns lead_time concept; authoritative record prevails."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "STALE-03", "family": "STALE", "rep": 0, "latency_s": 13.96, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"No, superseded by 2026-04-01 update which added exclusion of customers in notice period\",\n  \"owner_team\": \"finance\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"2025-01-01 definition replaced by refined 2026-04-01 version\"\n}\n```\n", "obj": {"answer": "No, superseded by 2026-04-01 update which added exclusion of customers in notice period", "owner_team": "finance", "source_id": null, "policy_id": null, "unanswerable": false, "stale": true, "note": "2025-01-01 definition replaced by refined 2026-04-01 version"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dim_no_gov", "item": "CONF-04", "family": "CONF", "rep": 0, "latency_s": 44.13, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Five years after contract ends (Legal definition D-802 from 2026-06-01)\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": \"D-802\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support record D-803 states seven years. Conflict unresolved in memory.\"\n}\n```\n", "obj": {"answer": "Five years after contract ends (Legal definition D-802 from 2026-06-01)", "owner_team": "legal", "source_id": "D-802", "policy_id": "D-802", "unanswerable": false, "stale": false, "note": "Support record D-803 states seven years. Conflict unresolved in memory."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "AUTH-05", "family": "AUTH", "rep": 0, "latency_s": 12.32, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"23 working days\",\n  \"owner_team\": \"Engineering\",\n  \"source_id\": \"N-401\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Engineering's figure prevails over Sales' competing entry per CP-ORG\"\n}\n```\n", "obj": {"answer": "23 working days", "owner_team": "Engineering", "source_id": "N-401", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Engineering's figure prevails over Sales' competing entry per CP-ORG"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "STALE-01", "family": "STALE", "rep": 1, "latency_s": 45.04, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"No. The authoritative company policy is five years after contract end, effective June 1, 2026.\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-802 is authoritative; D-803 is Support's non-authoritative view.\"\n}\n```\n", "obj": {"answer": "No. The authoritative company policy is five years after contract end, effective June 1, 2026.", "owner_team": "legal", "source_id": "D-802", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "D-802 is authoritative; D-803 is Support's non-authoritative view."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dim_no_gov", "item": "CONF-05", "family": "CONF", "rep": 1, "latency_s": 23.51, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Support (D-402) and Engineering (D-403) definitions differ; no company-wide resolution documented.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Support (D-402) and Engineering (D-403) definitions differ; no company-wide resolution documented."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "STALE-05", "family": "STALE", "rep": 0, "latency_s": 9.84, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Yes, Finance definition of logo churn is still in force, effective from 2025-01-01\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"FIN-POL-05\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Version 1 remains current as of today.\"\n}\n```\n", "obj": {"answer": "Yes, Finance definition of logo churn is still in force, effective from 2025-01-01", "owner_team": "Finance", "source_id": "FIN-POL-05", "policy_id": null, "unanswerable": false, "stale": false, "note": "Version 1 remains current as of today."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "bitemporal", "item": "PROV-02", "family": "PROV", "rep": 1, "latency_s": 11.25, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Sales team reported 12.4m EUR in SAL-DECK-Q1\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"N-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Record valid 2026-01-01 to 2026-04-30, now expired as of 2026-09-07\"\n}\n```\n", "obj": {"answer": "Sales team reported 12.4m EUR in SAL-DECK-Q1", "owner_team": "Sales", "source_id": "N-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "Record valid 2026-01-01 to 2026-04-30, now expired as of 2026-09-07"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "ROUTE-02", "family": "ROUTE", "rep": 1, "latency_s": 36.23, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Support\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Engineering D-403 also currently valid\"\n}\n```\n", "obj": {"answer": "Support", "owner_team": "Support", "source_id": "D-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "Engineering D-403 also currently valid"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "STALE-02", "family": "STALE", "rep": 0, "latency_s": 46.74, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"No. Support owns the severity 1 definition. The company uses Support's Version 2 (2026-01-15): 'an outage affecting any customer with no workaround, or any confirmed data loss.'\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"SUP-POL-03\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Engineering's definition is Engineering's view; Support owns it.\"\n}\n```\n", "obj": {"answer": "No. Support owns the severity 1 definition. The company uses Support's Version 2 (2026-01-15): 'an outage affecting any customer with no workaround, or any confirmed data loss.'", "owner_team": "Support", "source_id": "SUP-POL-03", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Engineering's definition is Engineering's view; Support owns it."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "vector_memory", "item": "HIST-02", "family": "HIST", "rep": 1, "latency_s": 13.53, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"customer records are kept for seven years after the contract ends\",\n  \"owner_team\": \"legal\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Policy superseded by 2026-06-01 legal entry: five years\"\n}\n```\n", "obj": {"answer": "customer records are kept for seven years after the contract ends", "owner_team": "legal", "source_id": null, "policy_id": null, "unanswerable": false, "stale": true, "note": "Policy superseded by 2026-06-01 legal entry: five years"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "HIST-04", "family": "HIST", "rep": 0, "latency_s": 38.33, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Version 2\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-403 also active with different definition\"\n}\n```\n", "obj": {"answer": "Version 2", "owner_team": "support", "source_id": "D-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "D-403 also active with different definition"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "HIST-04", "family": "HIST", "rep": 1, "latency_s": 12.26, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"an outage affecting any customer with no workaround, or any confirmed data loss\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-402 prevails over concurrent D-403 per ownership rule\"\n}\n```\n", "obj": {"answer": "an outage affecting any customer with no workaround, or any confirmed data loss", "owner_team": "Support", "source_id": "D-402", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "D-402 prevails over concurrent D-403 per ownership rule"}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "AUTH-03", "family": "AUTH", "rep": 0, "latency_s": 20.12, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"318\",\n  \"owner_team\": \"sales\",\n  \"source_id\": null,\n  \"policy_id\": \"2024-09-01\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales definition (live contract) applied as authoritative over support's ticket-based metric.\"\n}\n```\n", "obj": {"answer": "318", "owner_team": "sales", "source_id": null, "policy_id": "2024-09-01", "unanswerable": false, "stale": false, "note": "Sales definition (live contract) applied as authoritative over support's ticket-based metric."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "AUTH-05", "family": "AUTH", "rep": 0, "latency_s": 31.63, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"41 calendar days\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"N-402\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Measures purchase order to on-site delivery; Engineering reports different phase metric.\"\n}\n```\n", "obj": {"answer": "41 calendar days", "owner_team": "Sales", "source_id": "N-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "Measures purchase order to on-site delivery; Engineering reports different phase metric."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 1, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "GAP-01", "family": "GAP", "rep": 1, "latency_s": 13.74, "context_chars": 10621, "raw": "```json\n{\"answer\": null, \"owner_team\": \"Finance\", \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"Memory contains current ARR figures for 2026, not 2027 targets.\"}\n```\n", "obj": {"answer": null, "owner_team": "Finance", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains current ARR figures for 2026, not 2027 targets."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "graph_catalog", "item": "GAP-03", "family": "GAP", "rep": 0, "latency_s": 10.95, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Gross margin is not defined in the shared memory.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Gross margin is not defined in the shared memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "STALE-03", "family": "STALE", "rep": 0, "latency_s": 49.77, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"No. Finance ARR version 1 expired 2026-03-31, replaced by version 2.\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-101 expired and superseded. D-103 (sales v1) remains active.\"\n}\n```\n", "obj": {"answer": "No. Finance ARR version 1 expired 2026-03-31, replaced by version 2.", "owner_team": "finance", "source_id": "D-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "D-101 expired and superseded. D-103 (sales v1) remains active."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "bitemporal", "item": "STALE-02", "family": "STALE", "rep": 0, "latency_s": 31.44, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Yes, it is currently valid; however, Support's 2026-01-15 definition (D-402) also applies.\",\n  \"owner_team\": \"engineering\",\n  \"source_id\": \"D-403\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Two conflicting definitions are currently valid for severity 1 incidents.\"\n}\n```\n", "obj": {"answer": "Yes, it is currently valid; however, Support's 2026-01-15 definition (D-402) also applies.", "owner_team": "engineering", "source_id": "D-403", "policy_id": null, "unanswerable": false, "stale": false, "note": "Two conflicting definitions are currently valid for severity 1 incidents."}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "bitemporal", "item": "HIST-02", "family": "HIST", "rep": 0, "latency_s": 15.42, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"customer records are kept for seven years after the contract ends\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"D-801\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Valid on 2026-03-01 but superseded on 2026-06-01\"\n}\n```\n", "obj": {"answer": "customer records are kept for seven years after the contract ends", "owner_team": "Legal", "source_id": "D-801", "policy_id": null, "unanswerable": false, "stale": true, "note": "Valid on 2026-03-01 but superseded on 2026-06-01"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "HIST-03", "family": "HIST", "rep": 1, "latency_s": 16.71, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"an outage affecting more than one customer with no workaround\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"D-401\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-401 (Support owner) prevailed over concurrent D-403 per CP-ORG\"\n}\n```\n", "obj": {"answer": "an outage affecting more than one customer with no workaround", "owner_team": "Support", "source_id": "D-401", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "D-401 (Support owner) prevailed over concurrent D-403 per CP-ORG"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "CONF-03", "family": "CONF", "rep": 1, "latency_s": 20.61, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"Engineering's 23 working days (N-401) stands as the authoritative lead time measure.\",\n  \"owner_team\": \"engineering\",\n  \"source_id\": \"N-401\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Engineering owns the lead_time concept; its authoritative record prevails per CP-ORG.\"\n}\n```\n", "obj": {"answer": "Engineering's 23 working days (N-401) stands as the authoritative lead time measure.", "owner_team": "engineering", "source_id": "N-401", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Engineering owns the lead_time concept; its authoritative record prevails per CP-ORG."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "HIST-02", "family": "HIST", "rep": 0, "latency_s": 13.49, "context_chars": 10621, "raw": "```json\n{\"answer\": \"customer records are kept for seven years after the contract ends\", \"owner_team\": \"Legal\", \"source_id\": \"D-801\", \"policy_id\": null, \"unanswerable\": false, \"stale\": true, \"note\": \"D-801 superseded by D-802 on 2026-06-01\"}\n```\n", "obj": {"answer": "customer records are kept for seven years after the contract ends", "owner_team": "Legal", "source_id": "D-801", "policy_id": null, "unanswerable": false, "stale": true, "note": "D-801 superseded by D-802 on 2026-06-01"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "STALE-02", "family": "STALE", "rep": 1, "latency_s": 38.4, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"No. Support's definition (D-402) from 2026-01-15 supersedes it and defines severity 1 differently.\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"D-403\",\n  \"policy_id\": \"D-402\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Engineering's definition conflicts with Support's more recent 2026-01-15 definition, both currently valid.\"\n}\n```\n", "obj": {"answer": "No. Support's definition (D-402) from 2026-01-15 supersedes it and defines severity 1 differently.", "owner_team": "Support", "source_id": "D-403", "policy_id": "D-402", "unanswerable": false, "stale": true, "note": "Engineering's definition conflicts with Support's more recent 2026-01-15 definition, both currently valid."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dim_no_gov", "item": "CONF-03", "family": "CONF", "rep": 1, "latency_s": 26.55, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Both are valid; they measure different fulfillment phases\",\n  \"owner_team\": null,\n  \"source_id\": \"D-601, D-602\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Engineering: order→despatch (working days); Sales: PO→delivery (calendar days)\"\n}\n```\n", "obj": {"answer": "Both are valid; they measure different fulfillment phases", "owner_team": null, "source_id": "D-601, D-602", "policy_id": null, "unanswerable": false, "stale": false, "note": "Engineering: order→despatch (working days); Sales: PO→delivery (calendar days)"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 1, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "GAP-02", "family": "GAP", "rep": 1, "latency_s": 11.91, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": \"People\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory contains only engineering headcount growth plan for 2027, not current total.\"\n}\n```\n", "obj": {"answer": null, "owner_team": "People", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains only engineering headcount growth plan for 2027, not current total."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dimension", "item": "ROUTE-02", "family": "ROUTE", "rep": 0, "latency_s": 27.39, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"Support via the Support Lead, who owns the incident_severity_1 concept\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Concept ownership determines decision authority\"\n}\n```\n", "obj": {"answer": "Support via the Support Lead, who owns the incident_severity_1 concept", "owner_team": "support", "source_id": "D-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "Concept ownership determines decision authority"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "ROUTE-02", "family": "ROUTE", "rep": 1, "latency_s": 23.19, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Support team, accountable role: Support Lead\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"SUP-POL-03\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support owns severity 1 incident concept per governance catalogue\"\n}\n```\n", "obj": {"answer": "Support team, accountable role: Support Lead", "owner_team": "Support", "source_id": "SUP-POL-03", "policy_id": null, "unanswerable": false, "stale": false, "note": "Support owns severity 1 incident concept per governance catalogue"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "PROV-03", "family": "PROV", "rep": 0, "latency_s": 20.14, "context_chars": 10621, "raw": "```json\n{\"answer\": \"LEG-POL-04\", \"owner_team\": \"Legal\", \"source_id\": \"D-802\", \"policy_id\": \"CP-ORG\", \"unanswerable\": false, \"stale\": false, \"note\": \"Legal owns the concept; D-802 prevails over D-803 per CP-ORG\"}\n```\n", "obj": {"answer": "LEG-POL-04", "owner_team": "Legal", "source_id": "D-802", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Legal owns the concept; D-802 prevails over D-803 per CP-ORG"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "ROUTE-02", "family": "ROUTE", "rep": 0, "latency_s": 27.99, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Support\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"2026-01-15 (support) incident_severity_1\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support owns most recent definition; note engineering also defined severity 1 differently\"\n}\n```\n", "obj": {"answer": "Support", "owner_team": "support", "source_id": "2026-01-15 (support) incident_severity_1", "policy_id": null, "unanswerable": false, "stale": false, "note": "Support owns most recent definition; note engineering also defined severity 1 differently"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "STALE-03", "family": "STALE", "rep": 0, "latency_s": 25.06, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"No, version 1 is no longer in force (valid until 2026-03-31; version 2 took effect 2026-04-01)\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"V1 expired; V2 now authoritative\"\n}\n```\n", "obj": {"answer": "No, version 1 is no longer in force (valid until 2026-03-31; version 2 took effect 2026-04-01)", "owner_team": "finance", "source_id": "D-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "V1 expired; V2 now authoritative"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "bitemporal", "item": "CONF-03", "family": "CONF", "rep": 1, "latency_s": 51.84, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Both stand, measuring different lead time segments: Engineering reports 23 working days from confirmed order to despatch (D-601/N-401); Sales reports 41 calendar days from PO to on-site delivery (D-602/N-402).\",\n  \"owner_team\": null,\n  \"source_id\": \"D-601, D-602, N-401, N-402\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Both definitions valid; measure different process segments\"\n}\n```\n", "obj": {"answer": "Both stand, measuring different lead time segments: Engineering reports 23 working days from confirmed order to despatch (D-601/N-401); Sales reports 41 calendar days from PO to on-site delivery (D-602/N-402).", "owner_team": null, "source_id": "D-601, D-602, N-401, N-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "Both definitions valid; measure different process segments"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "CONF-01", "family": "CONF", "rep": 0, "latency_s": 20.6, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"9.8m EUR (Finance)\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"N-102\",\n  \"policy_id\": \"D-102\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance figure remains valid; Sales figure stale since 2026-04-30. D-102 definition governs.\"\n}\n```\n", "obj": {"answer": "9.8m EUR (Finance)", "owner_team": "Finance", "source_id": "N-102", "policy_id": "D-102", "unanswerable": false, "stale": false, "note": "Finance figure remains valid; Sales figure stale since 2026-04-30. D-102 definition governs."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "HIST-01", "family": "HIST", "rep": 1, "latency_s": 12.93, "context_chars": 6618, "raw": "```json\n{\"answer\": \"Version 1\", \"owner_team\": \"Finance\", \"source_id\": \"FIN-POL-04\", \"policy_id\": null, \"unanswerable\": false, \"stale\": true, \"note\": \"Version 1 in force 2025-01-01 to 2026-03-31; superseded by Version 2 on 2026-04-01\"}\n```\n", "obj": {"answer": "Version 1", "owner_team": "Finance", "source_id": "FIN-POL-04", "policy_id": null, "unanswerable": false, "stale": true, "note": "Version 1 in force 2025-01-01 to 2026-03-31; superseded by Version 2 on 2026-04-01"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "ROUTE-02", "family": "ROUTE", "rep": 0, "latency_s": 30.89, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Support - they own the incident_severity_1 policy definition recorded as SUP-POL-03\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Engineering has concurrent definition D-403; coordination may be needed\"\n}\n```\n", "obj": {"answer": "Support - they own the incident_severity_1 policy definition recorded as SUP-POL-03", "owner_team": "support", "source_id": "D-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "Engineering has concurrent definition D-403; coordination may be needed"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "AUTH-02", "family": "AUTH", "rep": 0, "latency_s": 16.77, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Both sales and support teams own competing definitions. Sales defines it as having a live contract (D-201, valid since 2024-09-01); support defines it as having opened a ticket in ninety days (D-202, valid since 2025-02-01).\",\n  \"owner_team\": null,\n  \"source_id\": \"D-201, D-202\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Two teams own conflicting definitions, both currently valid.\"\n}\n```\n", "obj": {"answer": "Both sales and support teams own competing definitions. Sales defines it as having a live contract (D-201, valid since 2024-09-01); support defines it as having opened a ticket in ninety days (D-202, valid since 2025-02-01).", "owner_team": null, "source_id": "D-201, D-202", "policy_id": null, "unanswerable": false, "stale": false, "note": "Two teams own conflicting definitions, both currently valid."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "STALE-01", "family": "STALE", "rep": 0, "latency_s": 16.24, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"No. Legal redefined the policy to five years effective 2026-06-01 (D-802). The Support wiki (D-803) still states seven years but that is superseded.\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": \"D-802\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Legal's policy (5 years) supersedes Support wiki (7 years) from 2026-06-01.\"\n}\n```\n", "obj": {"answer": "No. Legal redefined the policy to five years effective 2026-06-01 (D-802). The Support wiki (D-803) still states seven years but that is superseded.", "owner_team": "legal", "source_id": "D-802", "policy_id": "D-802", "unanswerable": false, "stale": true, "note": "Legal's policy (5 years) supersedes Support wiki (7 years) from 2026-06-01."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "wiki", "item": "GAP-03", "family": "GAP", "rep": 1, "latency_s": 8.0, "context_chars": 6618, "raw": "```json\n{\"answer\": null, \"owner_team\": null, \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"Gross margin not defined in shared memory records.\"}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Gross margin not defined in shared memory records."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "wiki", "item": "GAP-01", "family": "GAP", "rep": 1, "latency_s": 10.54, "context_chars": 6618, "raw": "```json\n{\"answer\": null, \"owner_team\": \"Finance\", \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"Memory contains 2026 ARR reports only, no 2027 target.\"}\n```\n", "obj": {"answer": null, "owner_team": "Finance", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains 2026 ARR reports only, no 2027 target."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "wiki", "item": "CONF-06", "family": "CONF", "rep": 1, "latency_s": 10.7, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"A person engaged under a services agreement with no employment relationship. People owns the definition; their record prevails per CP-ORG.\",\n  \"owner_team\": \"People\",\n  \"source_id\": \"PPL-POL-02\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"People (owner) definition supersedes Finance view per conflict rule.\"\n}\n```\n", "obj": {"answer": "A person engaged under a services agreement with no employment relationship. People owns the definition; their record prevails per CP-ORG.", "owner_team": "People", "source_id": "PPL-POL-02", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "People (owner) definition supersedes Finance view per conflict rule."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "HIST-01", "family": "HIST", "rep": 0, "latency_s": 38.73, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"committed subscription value recognised over the next twelve months, excluding one-off services\",\n  \"owner_team\": \"finance\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Finance definition from 2025-01-01; superseded on 2026-04-01\"\n}\n```\n", "obj": {"answer": "committed subscription value recognised over the next twelve months, excluding one-off services", "owner_team": "finance", "source_id": null, "policy_id": null, "unanswerable": false, "stale": true, "note": "Finance definition from 2025-01-01; superseded on 2026-04-01"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "CONF-04", "family": "CONF", "rep": 0, "latency_s": 14.76, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"5 years after the contract ends, per Legal\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal owns concept; both records valid 2026-06-01. Owning team prevails.\"\n}\n```\n", "obj": {"answer": "5 years after the contract ends, per Legal", "owner_team": "Legal", "source_id": "D-802", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Legal owns concept; both records valid 2026-06-01. Owning team prevails."}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "AUTH-03", "family": "AUTH", "rep": 1, "latency_s": 10.09, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"318\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"SAL-REP-JUL\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales owns the definition. Their count prevails over Support's 402.\"\n}\n```\n", "obj": {"answer": "318", "owner_team": "Sales", "source_id": "SAL-REP-JUL", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Sales owns the definition. Their count prevails over Support's 402."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "ROUTE-03", "family": "ROUTE", "rep": 1, "latency_s": 12.83, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"The General Counsel at Legal\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"LEG-POL-04\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal owns the definition; current policy requires five-year retention after contract end.\"\n}\n```\n", "obj": {"answer": "The General Counsel at Legal", "owner_team": "Legal", "source_id": "LEG-POL-04", "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal owns the definition; current policy requires five-year retention after contract end."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "ROUTE-04", "family": "ROUTE", "rep": 1, "latency_s": 7.54, "context_chars": 2694, "raw": "```json\n{\"answer\": \"The sales team\", \"owner_team\": \"sales\", \"source_id\": null, \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Sales team defined qualified_lead on 2025-10-01\"}\n```\n", "obj": {"answer": "The sales team", "owner_team": "sales", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales team defined qualified_lead on 2025-10-01"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "STALE-03", "family": "STALE", "rep": 0, "latency_s": 11.64, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"No. D-101 expired on 2026-03-31 and is no longer in force.\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Superseded by D-102 effective 2026-04-01\"\n}\n```\n", "obj": {"answer": "No. D-101 expired on 2026-03-31 and is no longer in force.", "owner_team": "Finance", "source_id": "D-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "Superseded by D-102 effective 2026-04-01"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dim_no_gov", "item": "GAP-04", "family": "GAP", "rep": 0, "latency_s": 10.13, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Net promoter score not found in the shared memory.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Net promoter score not found in the shared memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "PROV-02", "family": "PROV", "rep": 1, "latency_s": 24.38, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Sales team, document SAL-DECK-Q1\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"N-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales reported 12.4m EUR ARR for Q1 2026\"\n}\n```\n", "obj": {"answer": "Sales team, document SAL-DECK-Q1", "owner_team": "sales", "source_id": "N-101", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales reported 12.4m EUR ARR for Q1 2026"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "STALE-04", "family": "STALE", "rep": 0, "latency_s": 15.18, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Yes. A lead with a named budget holder and a stated timeframe.\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"D-701\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Definition current, valid since 2025-10-01 with no expiration set\"\n}\n```\n", "obj": {"answer": "Yes. A lead with a named budget holder and a stated timeframe.", "owner_team": "Sales", "source_id": "D-701", "policy_id": null, "unanswerable": false, "stale": false, "note": "Definition current, valid since 2025-10-01 with no expiration set"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "bitemporal", "item": "AUTH-03", "family": "AUTH", "rep": 0, "latency_s": 30.69, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"318\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"N-201\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales reports 318 (contract-based definition). Support reports 402 (activity-based).\"\n}\n```\n", "obj": {"answer": "318", "owner_team": "Sales", "source_id": "N-201", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales reports 318 (contract-based definition). Support reports 402 (activity-based)."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "PROV-04", "family": "PROV", "rep": 0, "latency_s": 18.75, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"An outage affecting any customer with no workaround, or any confirmed data loss\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"2026-01-15 incident_severity_1\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Most recent definition from support, superseded earlier engineering and support versions\"\n}\n```\n", "obj": {"answer": "An outage affecting any customer with no workaround, or any confirmed data loss", "owner_team": "support", "source_id": "2026-01-15 incident_severity_1", "policy_id": null, "unanswerable": false, "stale": false, "note": "Most recent definition from support, superseded earlier engineering and support versions"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "PROV-01", "family": "PROV", "rep": 1, "latency_s": 38.76, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"FIN-POL-07\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-102\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Authoritative definition from Finance, valid 2026-04-01 onwards\"\n}\n```\n", "obj": {"answer": "FIN-POL-07", "owner_team": "finance", "source_id": "D-102", "policy_id": null, "unanswerable": false, "stale": false, "note": "Authoritative definition from Finance, valid 2026-04-01 onwards"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "STALE-03", "family": "STALE", "rep": 0, "latency_s": 16.21, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"No, it expired on 2026-03-31\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-101 superseded by D-102 on 2026-04-01\"\n}\n```\n", "obj": {"answer": "No, it expired on 2026-03-31", "owner_team": "Finance", "source_id": "D-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "D-101 superseded by D-102 on 2026-04-01"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "wiki", "item": "CONF-04", "family": "CONF", "rep": 1, "latency_s": 11.64, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Five years after the contract ends. Legal's Version 2, in force from 2026-06-01, prevails over Support's Version 1.\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"LEG-POL-04\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal owns the definition. Version 2 supersedes Support's view.\"\n}\n```\n", "obj": {"answer": "Five years after the contract ends. Legal's Version 2, in force from 2026-06-01, prevails over Support's Version 1.", "owner_team": "Legal", "source_id": "LEG-POL-04", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Legal owns the definition. Version 2 supersedes Support's view."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "ROUTE-04", "family": "ROUTE", "rep": 0, "latency_s": 10.01, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Sales\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"D-701\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales team authored the current qualified lead definition.\"\n}\n```\n", "obj": {"answer": "Sales", "owner_team": "sales", "source_id": "D-701", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales team authored the current qualified lead definition."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "STALE-04", "family": "STALE", "rep": 1, "latency_s": 14.98, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Yes, the definition is in force: a lead with a named budget holder and a stated timeframe\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"D-701\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Valid from 2025-10-01, no expiration date, still current\"\n}\n```\n", "obj": {"answer": "Yes, the definition is in force: a lead with a named budget holder and a stated timeframe", "owner_team": "Sales", "source_id": "D-701", "policy_id": null, "unanswerable": false, "stale": false, "note": "Valid from 2025-10-01, no expiration date, still current"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dim_no_gov", "item": "PROV-04", "family": "PROV", "rep": 0, "latency_s": 22.33, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"SUP-POL-03\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-402 (version 2) in force from 2026-01-15; D-403 also current but version 1.\"\n}\n```\n", "obj": {"answer": "SUP-POL-03", "owner_team": "support", "source_id": "D-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "D-402 (version 2) in force from 2026-01-15; D-403 also current but version 1."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "AUTH-01", "family": "AUTH", "rep": 0, "latency_s": 29.76, "context_chars": 2694, "raw": "```json\n{\"answer\": \"9.8m EUR\", \"owner_team\": \"finance\", \"source_id\": null, \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Finance excludes one-off services and notice-period customers; Sales reports 12.4m EUR with different scope\"}\n```\n", "obj": {"answer": "9.8m EUR", "owner_team": "finance", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance excludes one-off services and notice-period customers; Sales reports 12.4m EUR with different scope"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "CONF-01", "family": "CONF", "rep": 0, "latency_s": 40.84, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Finance's 9.8m EUR stands; rule: CP-ORG\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"FIN-STMT-Q1\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance owns ARR definition; CP-ORG rule gives owner's record precedence\"\n}\n```\n", "obj": {"answer": "Finance's 9.8m EUR stands; rule: CP-ORG", "owner_team": "Finance", "source_id": "FIN-STMT-Q1", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Finance owns ARR definition; CP-ORG rule gives owner's record precedence"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "ROUTE-05", "family": "ROUTE", "rep": 0, "latency_s": 17.76, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Engineering Manager\",\n  \"owner_team\": \"Engineering\",\n  \"source_id\": \"ENG-POL-02\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Governance catalogue specifies Engineering Manager as accountable role\"\n}\n```\n", "obj": {"answer": "Engineering Manager", "owner_team": "Engineering", "source_id": "ENG-POL-02", "policy_id": null, "unanswerable": false, "stale": false, "note": "Governance catalogue specifies Engineering Manager as accountable role"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "ROUTE-01", "family": "ROUTE", "rep": 0, "latency_s": 39.48, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Finance (Head of Finance)\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"FIN-POL-07\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance owns arr; Head of Finance accountable\"\n}\n```\n", "obj": {"answer": "Finance (Head of Finance)", "owner_team": "Finance", "source_id": "FIN-POL-07", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance owns arr; Head of Finance accountable"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "ROUTE-01", "family": "ROUTE", "rep": 1, "latency_s": 16.4, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Head of Finance must agree. Finance owns the concept and is accountable through FIN-POL-07.\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"FIN-POL-07\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance owns ARR; Head of Finance is accountable role\"\n}\n```\n", "obj": {"answer": "Head of Finance must agree. Finance owns the concept and is accountable through FIN-POL-07.", "owner_team": "Finance", "source_id": "FIN-POL-07", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance owns ARR; Head of Finance is accountable role"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "AUTH-02", "family": "AUTH", "rep": 0, "latency_s": 27.01, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Sales\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"D-201\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support defined conflicting version (D-202) also valid\"\n}\n```\n", "obj": {"answer": "Sales", "owner_team": "Sales", "source_id": "D-201", "policy_id": null, "unanswerable": false, "stale": false, "note": "Support defined conflicting version (D-202) also valid"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "GAP-02", "family": "GAP", "rep": 1, "latency_s": 10.66, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": \"people\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Only headcount growth plan available, not current total\"\n}\n```\n", "obj": {"answer": null, "owner_team": "people", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Only headcount growth plan available, not current total"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "graph_catalog", "item": "HIST-05", "family": "HIST", "rep": 1, "latency_s": 21.54, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"No\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-101 was authoritative on 2025-12-31; D-102 superseded it 2026-04-01\"\n}\n```\n", "obj": {"answer": "No", "owner_team": "Finance", "source_id": "D-101", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "D-101 was authoritative on 2025-12-31; D-102 superseded it 2026-04-01"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "HIST-03", "family": "HIST", "rep": 1, "latency_s": 26.01, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Version 1\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"SUP-POL-01\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Support Version 1 applied 2025-03-01 to 2026-01-14, then superseded\"\n}\n```\n", "obj": {"answer": "Version 1", "owner_team": "Support", "source_id": "SUP-POL-01", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "Support Version 1 applied 2025-03-01 to 2026-01-14, then superseded"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "AUTH-03", "family": "AUTH", "rep": 0, "latency_s": 13.72, "context_chars": 6474, "raw": "```json\n{\"answer\": \"318\", \"owner_team\": \"sales\", \"source_id\": \"N-201\", \"policy_id\": \"D-201\", \"unanswerable\": false, \"stale\": false, \"note\": \"Sales definition (live contract) authoritative; Support reported 402 using different definition\"}\n```\n", "obj": {"answer": "318", "owner_team": "sales", "source_id": "N-201", "policy_id": "D-201", "unanswerable": false, "stale": false, "note": "Sales definition (live contract) authoritative; Support reported 402 using different definition"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "ROUTE-04", "family": "ROUTE", "rep": 0, "latency_s": 8.08, "context_chars": 6618, "raw": "```json\n{\"answer\": \"VP Sales\", \"owner_team\": \"Sales\", \"source_id\": \"SAL-POL-06\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Definition of qualified lead owned by Sales VP\"}\n```\n", "obj": {"answer": "VP Sales", "owner_team": "Sales", "source_id": "SAL-POL-06", "policy_id": null, "unanswerable": false, "stale": false, "note": "Definition of qualified lead owned by Sales VP"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "HIST-05", "family": "HIST", "rep": 0, "latency_s": 18.36, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"No\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"[2026-04-01]\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Notice period exclusion first appeared 2026-04-01; absent from prior definitions\"\n}\n```\n", "obj": {"answer": "No", "owner_team": "finance", "source_id": "[2026-04-01]", "policy_id": null, "unanswerable": false, "stale": false, "note": "Notice period exclusion first appeared 2026-04-01; absent from prior definitions"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "GAP-03", "family": "GAP", "rep": 1, "latency_s": 7.81, "context_chars": 2694, "raw": "```json\n{\"answer\": null, \"owner_team\": null, \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"Gross margin definition not found in shared memory.\"}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Gross margin definition not found in shared memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "wiki", "item": "PROV-04", "family": "PROV", "rep": 1, "latency_s": 14.1, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"SUP-POL-03\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"SUP-POL-03\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support Version 2 in force since 2026-01-15\"\n}\n```\n", "obj": {"answer": "SUP-POL-03", "owner_team": "Support", "source_id": "SUP-POL-03", "policy_id": null, "unanswerable": false, "stale": false, "note": "Support Version 2 in force since 2026-01-15"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "HIST-01", "family": "HIST", "rep": 0, "latency_s": 13.79, "context_chars": 6618, "raw": "```json\n{\"answer\": \"Version 1\", \"owner_team\": \"Finance\", \"source_id\": \"FIN-POL-04\", \"policy_id\": null, \"unanswerable\": false, \"stale\": true, \"note\": \"Version 1 was in force until 2026-03-31; superseded by Version 2.\"}\n```\n", "obj": {"answer": "Version 1", "owner_team": "Finance", "source_id": "FIN-POL-04", "policy_id": null, "unanswerable": false, "stale": true, "note": "Version 1 was in force until 2026-03-31; superseded by Version 2."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "AUTH-04", "family": "AUTH", "rep": 1, "latency_s": 18.47, "context_chars": 11191, "raw": "```json\n{\"answer\": \"Legal team (General Counsel)\", \"owner_team\": \"legal\", \"source_id\": \"D-802\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Legal owns data retention policy per concept ownership\"}\n```\n", "obj": {"answer": "Legal team (General Counsel)", "owner_team": "legal", "source_id": "D-802", "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal owns data retention policy per concept ownership"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "AUTH-04", "family": "AUTH", "rep": 1, "latency_s": 46.6, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Legal\",\n  \"owner_team\": \"legal\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal recorded data retention policy. Support also recorded on 2026-06-01, with conflicting value.\"\n}\n```\n", "obj": {"answer": "Legal", "owner_team": "legal", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal recorded data retention policy. Support also recorded on 2026-06-01, with conflicting value."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "CONF-03", "family": "CONF", "rep": 1, "latency_s": 19.82, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Engineering's record stands: 23 working days (N-401). Engineering owns the lead time concept per rule CP-ORG.\",\n  \"owner_team\": \"Engineering\",\n  \"source_id\": \"N-401\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales reports 41 calendar days (N-402), measuring customer PO to on-site delivery; Engineering measures order to despatch.\"\n}\n```\n", "obj": {"answer": "Engineering's record stands: 23 working days (N-401). Engineering owns the lead time concept per rule CP-ORG.", "owner_team": "Engineering", "source_id": "N-401", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Sales reports 41 calendar days (N-402), measuring customer PO to on-site delivery; Engineering measures order to despatch."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "CONF-01", "family": "CONF", "rep": 1, "latency_s": 46.57, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Finance's 9.8m EUR figure stands. Finance ARR definition excludes one-off services and notice-period customers; Sales includes one-off services.\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"2026-01-01\",\n  \"policy_id\": \"2026-04-01\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance ARR definition is more restrictive; Finance owns financial reporting authority.\"\n}\n```\n", "obj": {"answer": "Finance's 9.8m EUR figure stands. Finance ARR definition excludes one-off services and notice-period customers; Sales includes one-off services.", "owner_team": "Finance", "source_id": "2026-01-01", "policy_id": "2026-04-01", "unanswerable": false, "stale": false, "note": "Finance ARR definition is more restrictive; Finance owns financial reporting authority."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "CONF-06", "family": "CONF", "rep": 0, "latency_s": 11.27, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"a person engaged under a services agreement with no employment relationship\",\n  \"owner_team\": \"People\",\n  \"source_id\": \"D-501\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"People owns contractor. CP-ORG: owning team prevails when definitions conflict.\"\n}\n```\n", "obj": {"answer": "a person engaged under a services agreement with no employment relationship", "owner_team": "People", "source_id": "D-501", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "People owns contractor. CP-ORG: owning team prevails when definitions conflict."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "HIST-04", "family": "HIST", "rep": 0, "latency_s": 28.07, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"an outage affecting any customer with no workaround, or any confirmed data loss\",\n  \"owner_team\": \"support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-402 (2026-01-15) supersedes earlier D-403 (2025-11-01) by recency.\"\n}\n```\n", "obj": {"answer": "an outage affecting any customer with no workaround, or any confirmed data loss", "owner_team": "support", "source_id": "D-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "D-402 (2026-01-15) supersedes earlier D-403 (2025-11-01) by recency."}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "STALE-03", "family": "STALE", "rep": 1, "latency_s": 39.14, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Yes, the sales version (D-103) is in force\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"D-103\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance version D-101 expired 2026-03-31; sales version D-103 current\"\n}\n```\n", "obj": {"answer": "Yes, the sales version (D-103) is in force", "owner_team": "sales", "source_id": "D-103", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance version D-101 expired 2026-03-31; sales version D-103 current"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "vector_memory", "item": "HIST-04", "family": "HIST", "rep": 0, "latency_s": 14.7, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"an outage affecting any customer with no workaround, or any confirmed data loss\",\n  \"owner_team\": \"support\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Superseded by 2026-04-01 support record (value: 7).\"\n}\n```\n", "obj": {"answer": "an outage affecting any customer with no workaround, or any confirmed data loss", "owner_team": "support", "source_id": null, "policy_id": null, "unanswerable": false, "stale": true, "note": "Superseded by 2026-04-01 support record (value: 7)."}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "PROV-02", "family": "PROV", "rep": 0, "latency_s": 37.35, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Sales (SAL-DECK-Q1)\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"N-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Valid until 2026-04-30; finance reported 9.8m EUR instead.\"\n}\n```\n", "obj": {"answer": "Sales (SAL-DECK-Q1)", "owner_team": "sales", "source_id": "N-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "Valid until 2026-04-30; finance reported 9.8m EUR instead."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "GAP-02", "family": "GAP", "rep": 0, "latency_s": 11.52, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": \"People\",\n  \"source_id\": \"S-105\",\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory contains only engineering headcount plan, not total headcount.\"\n}\n```\n", "obj": {"answer": null, "owner_team": "People", "source_id": "S-105", "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains only engineering headcount plan, not total headcount."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "wiki", "item": "PROV-03", "family": "PROV", "rep": 0, "latency_s": 21.12, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"LEG-POL-04\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"LEG-POL-04\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal owns definition. LEG-POL-04 v2 from 2026-06-01 prevails over Support.\"\n}\n```\n", "obj": {"answer": "LEG-POL-04", "owner_team": "Legal", "source_id": "LEG-POL-04", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Legal owns definition. LEG-POL-04 v2 from 2026-06-01 prevails over Support."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "PROV-02", "family": "PROV", "rep": 0, "latency_s": 19.47, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"Sales team, document SAL-DECK-Q1\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"N-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Record expired 2026-04-30; currently stale.\"\n}\n```\n", "obj": {"answer": "Sales team, document SAL-DECK-Q1", "owner_team": "finance", "source_id": "N-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "Record expired 2026-04-30; currently stale."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "HIST-05", "family": "HIST", "rep": 1, "latency_s": 34.54, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"No\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"[2025-01-01]\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Finance ARR definition updated 2026-04-01 to exclude notice period\"\n}\n```\n", "obj": {"answer": "No", "owner_team": "finance", "source_id": "[2025-01-01]", "policy_id": null, "unanswerable": false, "stale": true, "note": "Finance ARR definition updated 2026-04-01 to exclude notice period"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "ROUTE-04", "family": "ROUTE", "rep": 1, "latency_s": 11.57, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Sales\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"D-701\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales team wrote the qualified lead definition in October 2025.\"\n}\n```\n", "obj": {"answer": "Sales", "owner_team": "Sales", "source_id": "D-701", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales team wrote the qualified lead definition in October 2025."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "AUTH-03", "family": "AUTH", "rep": 1, "latency_s": 20.75, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"318\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"N-201\",\n  \"policy_id\": \"D-201\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales reported 318; Support reported 402 using different definition. Contract-based definition authoritative.\"\n}\n```\n", "obj": {"answer": "318", "owner_team": "Sales", "source_id": "N-201", "policy_id": "D-201", "unanswerable": false, "stale": false, "note": "Sales reported 318; Support reported 402 using different definition. Contract-based definition authoritative."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "HIST-02", "family": "HIST", "rep": 1, "latency_s": 16.75, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"customer records are kept for seven years after the contract ends\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-801\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Record D-801 valid until 2026-05-31; superseded by D-802 on 2026-06-01\"\n}\n```\n", "obj": {"answer": "customer records are kept for seven years after the contract ends", "owner_team": "legal", "source_id": "D-801", "policy_id": null, "unanswerable": false, "stale": true, "note": "Record D-801 valid until 2026-05-31; superseded by D-802 on 2026-06-01"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "PROV-02", "family": "PROV", "rep": 1, "latency_s": 20.32, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"sales in SAL-DECK-Q1\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"N-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Validity ended 2026-04-30; superseded by N-102 reporting 9.8m EUR\"\n}\n```\n", "obj": {"answer": "sales in SAL-DECK-Q1", "owner_team": "finance", "source_id": "N-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "Validity ended 2026-04-30; superseded by N-102 reporting 9.8m EUR"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "CONF-02", "family": "CONF", "rep": 0, "latency_s": 13.9, "context_chars": 6618, "raw": "```json\n{\"answer\": \"318\", \"owner_team\": \"Sales\", \"source_id\": \"SAL-REP-JUL\", \"policy_id\": \"CP-ORG\", \"unanswerable\": false, \"stale\": false, \"note\": \"Sales owns definition; their report prevails per CP-ORG.\"}\n```\n", "obj": {"answer": "318", "owner_team": "Sales", "source_id": "SAL-REP-JUL", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Sales owns definition; their report prevails per CP-ORG."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "STALE-04", "family": "STALE", "rep": 0, "latency_s": 10.92, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Yes. A qualified lead is a lead with a named budget holder and a stated timeframe.\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"D-701\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Valid from 2025-10-01 with no end date.\"\n}\n```\n", "obj": {"answer": "Yes. A qualified lead is a lead with a named budget holder and a stated timeframe.", "owner_team": "sales", "source_id": "D-701", "policy_id": null, "unanswerable": false, "stale": false, "note": "Valid from 2025-10-01 with no end date."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "graph_catalog", "item": "GAP-04", "family": "GAP", "rep": 1, "latency_s": 12.23, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Net promoter score is not recorded in organizational memory.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Net promoter score is not recorded in organizational memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dimension", "item": "STALE-03", "family": "STALE", "rep": 1, "latency_s": 15.78, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"No, version 1 expired on 2026-03-31. Version 2 is now in force from 2026-04-01.\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-101 valid until 2026-03-31. Superseded by D-102 (version 2).\"\n}\n```\n", "obj": {"answer": "No, version 1 expired on 2026-03-31. Version 2 is now in force from 2026-04-01.", "owner_team": "finance", "source_id": "D-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "D-101 valid until 2026-03-31. Superseded by D-102 (version 2)."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "wiki", "item": "ROUTE-03", "family": "ROUTE", "rep": 0, "latency_s": 12.82, "context_chars": 6618, "raw": "```json\n{\"answer\": \"The General Counsel at Legal\", \"owner_team\": \"Legal\", \"source_id\": \"LEG-POL-04\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Legal owns customer data retention period definition.\"}\n```\n", "obj": {"answer": "The General Counsel at Legal", "owner_team": "Legal", "source_id": "LEG-POL-04", "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal owns customer data retention period definition."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "HIST-03", "family": "HIST", "rep": 0, "latency_s": 34.87, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"an outage breaching the uptime term of a signed contract\",\n  \"owner_team\": \"engineering\",\n  \"source_id\": \"2025-11-01\",\n  \"policy_id\": \"latest_timestamp\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Engineering definition; support defined differently on 2025-03-01\"\n}\n```\n", "obj": {"answer": "an outage breaching the uptime term of a signed contract", "owner_team": "engineering", "source_id": "2025-11-01", "policy_id": "latest_timestamp", "unanswerable": false, "stale": true, "note": "Engineering definition; support defined differently on 2025-03-01"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "CONF-04", "family": "CONF", "rep": 0, "latency_s": 41.68, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"5 years after contract ends\",\n  \"owner_team\": \"legal\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal 5yr vs Support 7yr disagreement, both 2026-06-01\"\n}\n```\n", "obj": {"answer": "5 years after contract ends", "owner_team": "legal", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal 5yr vs Support 7yr disagreement, both 2026-06-01"}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 1, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "STALE-03", "family": "STALE", "rep": 1, "latency_s": 16.72, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"No, it was superseded by an updated definition on 2026-04-01\",\n  \"owner_team\": \"finance\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"2025-01-01 definition updated 2026-04-01 to exclude customers in notice\"\n}\n```\n", "obj": {"answer": "No, it was superseded by an updated definition on 2026-04-01", "owner_team": "finance", "source_id": null, "policy_id": null, "unanswerable": false, "stale": true, "note": "2025-01-01 definition updated 2026-04-01 to exclude customers in notice"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dimension", "item": "AUTH-04", "family": "AUTH", "rep": 0, "latency_s": 32.12, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"General Counsel (Legal team)\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal team owns data retention concept; General Counsel role accountable.\"\n}\n```\n", "obj": {"answer": "General Counsel (Legal team)", "owner_team": "legal", "source_id": "D-802", "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal team owns data retention concept; General Counsel role accountable."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "CONF-05", "family": "CONF", "rep": 0, "latency_s": 39.35, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"an outage affecting any customer with no workaround, or any confirmed data loss\",\n  \"owner_team\": \"support\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support's 2026-01-15 definition most recent; differs from Engineering\"\n}\n```\n", "obj": {"answer": "an outage affecting any customer with no workaround, or any confirmed data loss", "owner_team": "support", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Support's 2026-01-15 definition most recent; differs from Engineering"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "GAP-01", "family": "GAP", "rep": 1, "latency_s": 9.21, "context_chars": 6474, "raw": "```json\n{\"answer\": null, \"owner_team\": null, \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"No 2027 ARR target found in shared memory\"}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "No 2027 ARR target found in shared memory"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dimension", "item": "ROUTE-04", "family": "ROUTE", "rep": 1, "latency_s": 11.64, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"The VP Sales (sales team)\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"D-701\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales owns qualified_lead concept; D-701 is authoritative definition\"\n}\n```\n", "obj": {"answer": "The VP Sales (sales team)", "owner_team": "sales", "source_id": "D-701", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales owns qualified_lead concept; D-701 is authoritative definition"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "GAP-01", "family": "GAP", "rep": 0, "latency_s": 11.26, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": \"Finance\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory contains 2026 ARR figures but no 2027 target.\"\n}\n```\n", "obj": {"answer": null, "owner_team": "Finance", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains 2026 ARR figures but no 2027 target."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "bitemporal", "item": "AUTH-06", "family": "AUTH", "rep": 1, "latency_s": 55.21, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Finance and People teams both define contractor differently; Finance's 2025 definition is more recent\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-501, D-502\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Two conflicting definitions; Finance's (2025) supersedes People's (2024) by recency\"\n}\n```\n", "obj": {"answer": "Finance and People teams both define contractor differently; Finance's 2025 definition is more recent", "owner_team": "Finance", "source_id": "D-501, D-502", "policy_id": null, "unanswerable": false, "stale": false, "note": "Two conflicting definitions; Finance's (2025) supersedes People's (2024) by recency"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "CONF-03", "family": "CONF", "rep": 0, "latency_s": 45.14, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Both measures valid; different scopes: internal vs customer-facing\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Both measures valid; different scopes: internal vs customer-facing"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "AUTH-05", "family": "AUTH", "rep": 1, "latency_s": 32.16, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"41 calendar days\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"N-402\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales definition: from customer purchase order to on-site delivery\"\n}\n```\n", "obj": {"answer": "41 calendar days", "owner_team": "sales", "source_id": "N-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales definition: from customer purchase order to on-site delivery"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 1, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "CONF-01", "family": "CONF", "rep": 0, "latency_s": 18.7, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"9.8m EUR from Finance (N-102) stands\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"N-102\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Finance owns ARR. Sales figure N-101 expired 2026-04-30. Finance authoritative.\"\n}\n```\n", "obj": {"answer": "9.8m EUR from Finance (N-102) stands", "owner_team": "finance", "source_id": "N-102", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "Finance owns ARR. Sales figure N-101 expired 2026-04-30. Finance authoritative."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "GAP-02", "family": "GAP", "rep": 0, "latency_s": 11.46, "context_chars": 2694, "raw": "```json\n{\"answer\": null, \"owner_team\": \"people\", \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"Memory contains only salary bands and growth plan, no current headcount.\"}\n```\n", "obj": {"answer": null, "owner_team": "people", "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains only salary bands and growth plan, no current headcount."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "bitemporal", "item": "HIST-05", "family": "HIST", "rep": 0, "latency_s": 16.93, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"No. The valid definition (D-101) on 2025-12-31 excluded only one-off services, not customers in notice period.\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-101 superseded by D-102 on 2026-04-01, which added notice period exclusion.\"\n}\n```\n", "obj": {"answer": "No. The valid definition (D-101) on 2025-12-31 excluded only one-off services, not customers in notice period.", "owner_team": "Finance", "source_id": "D-101", "policy_id": null, "unanswerable": false, "stale": true, "note": "D-101 superseded by D-102 on 2026-04-01, which added notice period exclusion."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "CONF-05", "family": "CONF", "rep": 1, "latency_s": 13.35, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"An outage affecting any customer with no workaround, or any confirmed data loss.\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"SUP-POL-03\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support owns the definition. Version 2 applies since 2026-01-15.\"\n}\n```\n", "obj": {"answer": "An outage affecting any customer with no workaround, or any confirmed data loss.", "owner_team": "Support", "source_id": "SUP-POL-03", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Support owns the definition. Version 2 applies since 2026-01-15."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "STALE-05", "family": "STALE", "rep": 0, "latency_s": 9.93, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory contains finance churn definition, not logo churn specifically\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains finance churn definition, not logo churn specifically"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "bitemporal", "item": "AUTH-06", "family": "AUTH", "rep": 0, "latency_s": 29.81, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Finance\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"D-502\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance definition (D-502) most recent; People team also has earlier definition (D-501)\"\n}\n```\n", "obj": {"answer": "Finance", "owner_team": "Finance", "source_id": "D-502", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance definition (D-502) most recent; People team also has earlier definition (D-501)"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "ROUTE-03", "family": "ROUTE", "rep": 0, "latency_s": 13.14, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"The Legal team (General Counsel). They own the data_retention concept.\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal is concept owner. Current policy: 5 years post-contract.\"\n}\n```\n", "obj": {"answer": "The Legal team (General Counsel). They own the data_retention concept.", "owner_team": "legal", "source_id": "D-802", "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal is concept owner. Current policy: 5 years post-contract."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "HIST-05", "family": "HIST", "rep": 1, "latency_s": 17.31, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"No. The ARR definition valid on 2025-12-31 (D-101) excluded one-off services but did not exclude customers in notice period. The exclusion was added in D-102 v2, effective 2026-04-01.\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": \"FIN-POL-04\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Definition superseded by D-102 on 2026-04-01 with notice period exclusion.\"\n}\n```\n", "obj": {"answer": "No. The ARR definition valid on 2025-12-31 (D-101) excluded one-off services but did not exclude customers in notice period. The exclusion was added in D-102 v2, effective 2026-04-01.", "owner_team": "finance", "source_id": "D-101", "policy_id": "FIN-POL-04", "unanswerable": false, "stale": true, "note": "Definition superseded by D-102 on 2026-04-01 with notice period exclusion."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "CONF-05", "family": "CONF", "rep": 1, "latency_s": 65.98, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Support's definition applies: an outage affecting any customer with no workaround, or any confirmed data loss\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Engineering defines differently (D-403); both definitions currently valid, unresolved\"\n}\n```\n", "obj": {"answer": "Support's definition applies: an outage affecting any customer with no workaround, or any confirmed data loss", "owner_team": "Support", "source_id": "D-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "Engineering defines differently (D-403); both definitions currently valid, unresolved"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "HIST-03", "family": "HIST", "rep": 1, "latency_s": 23.03, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"an outage breaching the uptime term of a signed contract\",\n  \"owner_team\": \"engineering\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Most recent definition before 2025-12-01 from engineering on 2025-11-01.\"\n}\n```\n", "obj": {"answer": "an outage breaching the uptime term of a signed contract", "owner_team": "engineering", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Most recent definition before 2025-12-01 from engineering on 2025-11-01."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "STALE-02", "family": "STALE", "rep": 0, "latency_s": 33.68, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"No. The company uses Support's definition (as owner): 'an outage affecting any customer with no workaround, or any confirmed data loss'.\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"D-403\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Engineering's definition retained as their view; Support owns concept.\"\n}\n```\n", "obj": {"answer": "No. The company uses Support's definition (as owner): 'an outage affecting any customer with no workaround, or any confirmed data loss'.", "owner_team": "Support", "source_id": "D-403", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Engineering's definition retained as their view; Support owns concept."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "graph_catalog", "item": "AUTH-06", "family": "AUTH", "rep": 1, "latency_s": 21.0, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"People\",\n  \"owner_team\": \"People\",\n  \"source_id\": \"PPL-POL-02\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Contractor owned by People; Head of People is accountable; authoritative document PPL-POL-02\"\n}\n```\n", "obj": {"answer": "People", "owner_team": "People", "source_id": "PPL-POL-02", "policy_id": null, "unanswerable": false, "stale": false, "note": "Contractor owned by People; Head of People is accountable; authoritative document PPL-POL-02"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "HIST-02", "family": "HIST", "rep": 0, "latency_s": 16.43, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"customer records are kept for seven years after the contract ends\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-801\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"D-801 was valid on 2026-03-01 but superseded by D-802 from 2026-06-01.\"\n}\n```\n", "obj": {"answer": "customer records are kept for seven years after the contract ends", "owner_team": "legal", "source_id": "D-801", "policy_id": null, "unanswerable": false, "stale": true, "note": "D-801 was valid on 2026-03-01 but superseded by D-802 from 2026-06-01."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "ROUTE-03", "family": "ROUTE", "rep": 0, "latency_s": 27.07, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Legal team\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": \"D-802\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal owns retention policy; Support definition conflicts on years\"\n}\n```\n", "obj": {"answer": "Legal team", "owner_team": "Legal", "source_id": "D-802", "policy_id": "D-802", "unanswerable": false, "stale": false, "note": "Legal owns retention policy; Support definition conflicts on years"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "CONF-02", "family": "CONF", "rep": 0, "latency_s": 12.33, "context_chars": 10621, "raw": "```json\n{\"answer\": \"318\", \"owner_team\": \"Sales\", \"source_id\": \"N-201\", \"policy_id\": \"CP-ORG\", \"unanswerable\": false, \"stale\": false, \"note\": \"Sales owns active_customer; owner's record prevails on same date.\"}\n```\n", "obj": {"answer": "318", "owner_team": "Sales", "source_id": "N-201", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Sales owns active_customer; owner's record prevails on same date."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "ROUTE-05", "family": "ROUTE", "rep": 1, "latency_s": 15.5, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"Engineering Manager\",\n  \"owner_team\": \"engineering\",\n  \"source_id\": \"D-601\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Engineering Manager owns the lead_time concept definition\"\n}\n```\n", "obj": {"answer": "Engineering Manager", "owner_team": "engineering", "source_id": "D-601", "policy_id": null, "unanswerable": false, "stale": false, "note": "Engineering Manager owns the lead_time concept definition"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "HIST-04", "family": "HIST", "rep": 1, "latency_s": 63.03, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"an outage affecting any customer with no workaround, or any confirmed data loss\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-403 from Engineering also valid; two definitions coexisted\"\n}\n```\n", "obj": {"answer": "an outage affecting any customer with no workaround, or any confirmed data loss", "owner_team": "Support", "source_id": "D-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "D-403 from Engineering also valid; two definitions coexisted"}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "STALE-02", "family": "STALE", "rep": 0, "latency_s": 32.95, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"No, Support's 2026-01-15 definition superseded it.\",\n  \"owner_team\": \"support\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Engineering definition was superseded by Support's later definition.\"\n}\n```\n", "obj": {"answer": "No, Support's 2026-01-15 definition superseded it.", "owner_team": "support", "source_id": null, "policy_id": null, "unanswerable": false, "stale": true, "note": "Engineering definition was superseded by Support's later definition."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "wiki", "item": "CONF-02", "family": "CONF", "rep": 1, "latency_s": 11.72, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Sales' 318 active customers stands\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"SAL-REP-JUL\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales owns the definition; CP-ORG resolves the disagreement for same-date reports.\"\n}\n```\n", "obj": {"answer": "Sales' 318 active customers stands", "owner_team": "Sales", "source_id": "SAL-REP-JUL", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Sales owns the definition; CP-ORG resolves the disagreement for same-date reports."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "STALE-05", "family": "STALE", "rep": 0, "latency_s": 10.15, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"No record of Finance definition for 'logo churn' in memory.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "No record of Finance definition for 'logo churn' in memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "wiki", "item": "GAP-04", "family": "GAP", "rep": 0, "latency_s": 8.38, "context_chars": 6618, "raw": "```json\n{\"answer\": null, \"owner_team\": null, \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"Net promoter score not mentioned in shared memory.\"}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Net promoter score not mentioned in shared memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "bitemporal", "item": "STALE-05", "family": "STALE", "rep": 0, "latency_s": 9.56, "context_chars": 8932, "raw": "```json\n{\"answer\": \"Yes, the Finance definition of logo churn (customers whose contract ended in the period and was not renewed within thirty days) is still in force.\", \"owner_team\": \"Finance\", \"source_id\": \"D-301\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Valid since 2025-01-01 with no end date (invalid_at: open)\"}\n```\n", "obj": {"answer": "Yes, the Finance definition of logo churn (customers whose contract ended in the period and was not renewed within thirty days) is still in force.", "owner_team": "Finance", "source_id": "D-301", "policy_id": null, "unanswerable": false, "stale": false, "note": "Valid since 2025-01-01 with no end date (invalid_at: open)"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "wiki", "item": "ROUTE-01", "family": "ROUTE", "rep": 0, "latency_s": 23.93, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Head of Finance must agree. They own the definition of annual recurring revenue.\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"FIN-POL-07\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Definition ownership vested in Head of Finance per org structure.\"\n}\n```\n", "obj": {"answer": "Head of Finance must agree. They own the definition of annual recurring revenue.", "owner_team": "Finance", "source_id": "FIN-POL-07", "policy_id": null, "unanswerable": false, "stale": false, "note": "Definition ownership vested in Head of Finance per org structure."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "HIST-04", "family": "HIST", "rep": 0, "latency_s": 11.44, "context_chars": 6618, "raw": "```json\n{\"answer\": \"Version 2: an outage affecting any customer with no workaround, or any confirmed data loss.\", \"owner_team\": \"Support\", \"source_id\": \"SUP-POL-03\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Version 2 in force from 2026-01-15. Support owns severity 1 definition.\"}\n```\n", "obj": {"answer": "Version 2: an outage affecting any customer with no workaround, or any confirmed data loss.", "owner_team": "Support", "source_id": "SUP-POL-03", "policy_id": null, "unanswerable": false, "stale": false, "note": "Version 2 in force from 2026-01-15. Support owns severity 1 definition."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "CONF-06", "family": "CONF", "rep": 0, "latency_s": 18.07, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory records two contractor definitions with no precedence rule specified.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory records two contractor definitions with no precedence rule specified."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "ROUTE-04", "family": "ROUTE", "rep": 1, "latency_s": 8.23, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"The sales team must be asked to change the definition of a qualified lead\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"D-701\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales team recorded the current definition via SAL-POL-06.\"\n}\n```\n", "obj": {"answer": "The sales team must be asked to change the definition of a qualified lead", "owner_team": "sales", "source_id": "D-701", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales team recorded the current definition via SAL-POL-06."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "PROV-01", "family": "PROV", "rep": 0, "latency_s": 29.31, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"FIN-POL-07\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-102\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance version 2, current from 2026-04-01. Sales definition D-103 also active.\"\n}\n```\n", "obj": {"answer": "FIN-POL-07", "owner_team": "finance", "source_id": "D-102", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance version 2, current from 2026-04-01. Sales definition D-103 also active."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "HIST-02", "family": "HIST", "rep": 0, "latency_s": 17.19, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"customer records are kept for seven years after the contract ends\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"LEG-POL-01\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Version 1 in effect; superseded 2026-06-01 by Version 2\"\n}\n```\n", "obj": {"answer": "customer records are kept for seven years after the contract ends", "owner_team": "Legal", "source_id": "LEG-POL-01", "policy_id": null, "unanswerable": false, "stale": true, "note": "Version 1 in effect; superseded 2026-06-01 by Version 2"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "AUTH-01", "family": "AUTH", "rep": 0, "latency_s": 12.47, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"9.8m EUR from the finance team\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"N-102\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance's authoritative figure valid from 2026-01-01 onwards.\"\n}\n```\n", "obj": {"answer": "9.8m EUR from the finance team", "owner_team": "finance", "source_id": "N-102", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance's authoritative figure valid from 2026-01-01 onwards."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "PROV-04", "family": "PROV", "rep": 1, "latency_s": 40.05, "context_chars": 10621, "raw": "```json\n{\"answer\": \"SUP-POL-03\", \"owner_team\": \"Support\", \"source_id\": \"D-402\", \"policy_id\": \"CP-ORG\", \"unanswerable\": false, \"stale\": false, \"note\": \"D-402 (Support team owner) in SUP-POL-03 prevails over D-403 (Engineering) per CP-ORG.\"}\n```\n", "obj": {"answer": "SUP-POL-03", "owner_team": "Support", "source_id": "D-402", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "D-402 (Support team owner) in SUP-POL-03 prevails over D-403 (Engineering) per CP-ORG."}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "CONF-04", "family": "CONF", "rep": 1, "latency_s": 12.57, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"Customer records are kept for five years after the contract ends. Legal's definition (D-802) prevails as the concept owner.\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal owns data_retention. D-802 supersedes conflicting Support record D-803.\"\n}\n```\n", "obj": {"answer": "Customer records are kept for five years after the contract ends. Legal's definition (D-802) prevails as the concept owner.", "owner_team": "legal", "source_id": "D-802", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Legal owns data_retention. D-802 supersedes conflicting Support record D-803."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "STALE-04", "family": "STALE", "rep": 1, "latency_s": 12.22, "context_chars": 2694, "raw": "```json\n{\"answer\": \"Yes. A lead with a named budget holder and a stated timeframe.\", \"owner_team\": \"sales\", \"source_id\": \"2025-10-01\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Definition established 2025-10-01 with no superseding record\"}\n```\n", "obj": {"answer": "Yes. A lead with a named budget holder and a stated timeframe.", "owner_team": "sales", "source_id": "2025-10-01", "policy_id": null, "unanswerable": false, "stale": false, "note": "Definition established 2025-10-01 with no superseding record"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "vector_memory", "item": "STALE-02", "family": "STALE", "rep": 1, "latency_s": 11.92, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"No. Engineering's 2025-11-01 definition was superseded by Support's 2026-01-15 definition: 'an outage affecting any customer with no workaround, or any confirmed data loss'.\",\n  \"owner_team\": \"support\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Engineering definition superseded by Support update 2026-01-15.\"\n}\n```\n", "obj": {"answer": "No. Engineering's 2025-11-01 definition was superseded by Support's 2026-01-15 definition: 'an outage affecting any customer with no workaround, or any confirmed data loss'.", "owner_team": "support", "source_id": null, "policy_id": null, "unanswerable": false, "stale": true, "note": "Engineering definition superseded by Support update 2026-01-15."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "vector_memory", "item": "PROV-02", "family": "PROV", "rep": 0, "latency_s": 14.15, "context_chars": 2694, "raw": "```json\n{\"answer\": \"sales\", \"owner_team\": \"sales\", \"source_id\": null, \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Sales team reported 12.4m EUR on 2026-01-01. Document not identified.\"}\n```\n", "obj": {"answer": "sales", "owner_team": "sales", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales team reported 12.4m EUR on 2026-01-01. Document not identified."}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "ROUTE-05", "family": "ROUTE", "rep": 0, "latency_s": 48.89, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Sales\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"D-602\",\n  \"policy_id\": \"D-602\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Two definitions exist; using most recent Sales definition\"\n}\n```\n", "obj": {"answer": "Sales", "owner_team": "Sales", "source_id": "D-602", "policy_id": "D-602", "unanswerable": false, "stale": false, "note": "Two definitions exist; using most recent Sales definition"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "HIST-01", "family": "HIST", "rep": 1, "latency_s": 21.22, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"1\",\n  \"owner_team\": \"finance\",\n  \"source_id\": \"D-101\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Conflicting definitions existed; D-101 (finance) prevailed per CP-ORG\"\n}\n```\n", "obj": {"answer": "1", "owner_team": "finance", "source_id": "D-101", "policy_id": "CP-ORG", "unanswerable": false, "stale": true, "note": "Conflicting definitions existed; D-101 (finance) prevailed per CP-ORG"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "AUTH-02", "family": "AUTH", "rep": 1, "latency_s": 32.4, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"Sales\",\n  \"owner_team\": \"Sales\",\n  \"source_id\": \"SAL-POL-02\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales owns active_customer per governance catalogue; authoritative document SAL-POL-02\"\n}\n```\n", "obj": {"answer": "Sales", "owner_team": "Sales", "source_id": "SAL-POL-02", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales owns active_customer per governance catalogue; authoritative document SAL-POL-02"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "STALE-01", "family": "STALE", "rep": 1, "latency_s": 19.44, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"No. Legal updated the policy to five years after contract end (2026-06-01), superseding the prior seven-year position.\",\n  \"owner_team\": \"legal\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Legal and Support records conflict; Legal is authoritative on data retention policy.\"\n}\n```\n", "obj": {"answer": "No. Legal updated the policy to five years after contract end (2026-06-01), superseding the prior seven-year position.", "owner_team": "legal", "source_id": null, "policy_id": null, "unanswerable": false, "stale": true, "note": "Legal and Support records conflict; Legal is authoritative on data retention policy."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "bitemporal", "item": "GAP-03", "family": "GAP", "rep": 0, "latency_s": 12.52, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Gross margin is not defined in the shared memory of Calder Instruments.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Gross margin is not defined in the shared memory of Calder Instruments."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "vector_memory", "item": "HIST-04", "family": "HIST", "rep": 1, "latency_s": 25.9, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"an outage affecting any customer with no workaround, or any confirmed data loss\",\n  \"owner_team\": \"support\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Most recent definition from support, 2026-01-15\"\n}\n```\n", "obj": {"answer": "an outage affecting any customer with no workaround, or any confirmed data loss", "owner_team": "support", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Most recent definition from support, 2026-01-15"}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "STALE-01", "family": "STALE", "rep": 0, "latency_s": 23.87, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Support's position is seven years (D-803, valid from 2026-06-01), but conflicts with Legal's current policy of five years (D-802, same validity date). Not a unified company position.\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"D-803\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Conflicting retention policies from Support and Legal, both current since 2026-06-01\"\n}\n```\n", "obj": {"answer": "Support's position is seven years (D-803, valid from 2026-06-01), but conflicts with Legal's current policy of five years (D-802, same validity date). Not a unified company position.", "owner_team": "Support", "source_id": "D-803", "policy_id": null, "unanswerable": false, "stale": false, "note": "Conflicting retention policies from Support and Legal, both current since 2026-06-01"}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dim_no_gov", "item": "CONF-05", "family": "CONF", "rep": 0, "latency_s": 46.91, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Conflicting definitions: Support D-402 vs Engineering D-403, both active\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Conflicting definitions: Support D-402 vs Engineering D-403, both active"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "AUTH-06", "family": "AUTH", "rep": 0, "latency_s": 17.34, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"People, specifically the Head of People\",\n  \"owner_team\": \"People\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Stated in 'Who owns what' section of shared memory\"\n}\n```\n", "obj": {"answer": "People, specifically the Head of People", "owner_team": "People", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Stated in 'Who owns what' section of shared memory"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "PROV-04", "family": "PROV", "rep": 0, "latency_s": 10.89, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"SUP-POL-03\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"SUP-POL-03\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support owns definition; version 2 in force from 2026-01-15\"\n}\n```\n", "obj": {"answer": "SUP-POL-03", "owner_team": "Support", "source_id": "SUP-POL-03", "policy_id": null, "unanswerable": false, "stale": false, "note": "Support owns definition; version 2 in force from 2026-01-15"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "PROV-04", "family": "PROV", "rep": 0, "latency_s": 43.65, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"SUP-POL-03\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": \"D-402\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-403 in ENG-WIKI-08 also currently defines severity 1\"\n}\n```\n", "obj": {"answer": "SUP-POL-03", "owner_team": "Support", "source_id": "D-402", "policy_id": "D-402", "unanswerable": false, "stale": false, "note": "D-403 in ENG-WIKI-08 also currently defines severity 1"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "CONF-06", "family": "CONF", "rep": 1, "latency_s": 12.63, "context_chars": 11191, "raw": "```json\n{\n  \"answer\": \"A person engaged under a services agreement with no employment relationship\",\n  \"owner_team\": \"people\",\n  \"source_id\": \"D-501\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"People owns contractor concept; their authoritative definition prevails.\"\n}\n```\n", "obj": {"answer": "A person engaged under a services agreement with no employment relationship", "owner_team": "people", "source_id": "D-501", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "People owns contractor concept; their authoritative definition prevails."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "ROUTE-02", "family": "ROUTE", "rep": 1, "latency_s": 17.48, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Support's decision. The Support Lead specifically owns the definition of severity 1 incidents.\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"SUP-POL-03\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support Lead owns severity 1 incident definition per shared memory.\"\n}\n```\n", "obj": {"answer": "Support's decision. The Support Lead specifically owns the definition of severity 1 incidents.", "owner_team": "Support", "source_id": "SUP-POL-03", "policy_id": null, "unanswerable": false, "stale": false, "note": "Support Lead owns severity 1 incident definition per shared memory."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "STALE-01", "family": "STALE", "rep": 0, "latency_s": 15.29, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"No. Legal updated the policy to five years after contract ends on 2026-06-01. Support's seven-year record is superseded.\",\n  \"owner_team\": \"legal\",\n  \"source_id\": null,\n  \"policy_id\": \"2026-06-01 legal data_retention update\",\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Support record conflicts with later legal update; legal controls policy.\"\n}\n```\n", "obj": {"answer": "No. Legal updated the policy to five years after contract ends on 2026-06-01. Support's seven-year record is superseded.", "owner_team": "legal", "source_id": null, "policy_id": "2026-06-01 legal data_retention update", "unanswerable": false, "stale": true, "note": "Support record conflicts with later legal update; legal controls policy."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "graph_catalog", "item": "CONF-04", "family": "CONF", "rep": 1, "latency_s": 17.0, "context_chars": 10621, "raw": "```json\n{\"answer\": \"Customer records are kept for five years after the contract ends\", \"owner_team\": \"Legal\", \"source_id\": \"D-802\", \"policy_id\": \"CP-ORG\", \"unanswerable\": false, \"stale\": false, \"note\": \"Legal owns data_retention; prevails per CP-ORG despite Support disagreement.\"}\n```\n", "obj": {"answer": "Customer records are kept for five years after the contract ends", "owner_team": "Legal", "source_id": "D-802", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Legal owns data_retention; prevails per CP-ORG despite Support disagreement."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "PROV-03", "family": "PROV", "rep": 1, "latency_s": 25.73, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"LEG-POL-04\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal-owned concept; CP-ORG applied to resolve D-802 vs D-803 conflict\"\n}\n```\n", "obj": {"answer": "LEG-POL-04", "owner_team": "Legal", "source_id": "D-802", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Legal-owned concept; CP-ORG applied to resolve D-802 vs D-803 conflict"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "GAP-04", "family": "GAP", "rep": 1, "latency_s": 8.97, "context_chars": 6618, "raw": "```json\n{\"answer\": null, \"owner_team\": null, \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"Net promoter score not mentioned in shared memory.\"}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Net promoter score not mentioned in shared memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "bitemporal", "item": "ROUTE-02", "family": "ROUTE", "rep": 0, "latency_s": 70.99, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"Not specified. Support and Engineering have competing definitions.\",\n  \"owner_team\": null,\n  \"source_id\": \"D-402\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Support defined D-402 (2026-01-15, current); Engineering's D-403 also valid.\"\n}\n```\n", "obj": {"answer": "Not specified. Support and Engineering have competing definitions.", "owner_team": null, "source_id": "D-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "Support defined D-402 (2026-01-15, current); Engineering's D-403 also valid."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "PROV-03", "family": "PROV", "rep": 1, "latency_s": 27.66, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"LEG-POL-04\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"D-802\",\n  \"policy_id\": \"D-802\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal policy 5 years supersedes old 7-year rule; support wiki conflicts\"\n}\n```\n", "obj": {"answer": "LEG-POL-04", "owner_team": "legal", "source_id": "D-802", "policy_id": "D-802", "unanswerable": false, "stale": false, "note": "Legal policy 5 years supersedes old 7-year rule; support wiki conflicts"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "HIST-04", "family": "HIST", "rep": 1, "latency_s": 12.3, "context_chars": 6618, "raw": "```json\n{\"answer\": \"Version 2\", \"owner_team\": \"Support\", \"source_id\": \"SUP-POL-03\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Version 2 in force from 2026-01-15, still current\"}\n```\n", "obj": {"answer": "Version 2", "owner_team": "Support", "source_id": "SUP-POL-03", "policy_id": null, "unanswerable": false, "stale": false, "note": "Version 2 in force from 2026-01-15, still current"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "CONF-03", "family": "CONF", "rep": 0, "latency_s": 11.0, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Engineering's record stands. The order lead time is owned by Engineering; per CP-ORG, the owning team's record prevails in disagreements.\",\n  \"owner_team\": \"Engineering\",\n  \"source_id\": \"ENG-REP-JUN\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Engineering owns order lead time definition; their report prevails over Sales.\"\n}\n```\n", "obj": {"answer": "Engineering's record stands. The order lead time is owned by Engineering; per CP-ORG, the owning team's record prevails in disagreements.", "owner_team": "Engineering", "source_id": "ENG-REP-JUN", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Engineering owns order lead time definition; their report prevails over Sales."}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "AUTH-01", "family": "AUTH", "rep": 1, "latency_s": 41.81, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"9.8m EUR\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"N-102\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Finance owns ARR. N-102 (9.8m) selected over N-101 (12.4m) by CP-ORG rule.\"\n}\n```\n", "obj": {"answer": "9.8m EUR", "owner_team": "Finance", "source_id": "N-102", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "Finance owns ARR. N-102 (9.8m) selected over N-101 (12.4m) by CP-ORG rule."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "HIST-01", "family": "HIST", "rep": 0, "latency_s": 35.96, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"version 1\",\n  \"owner_team\": null,\n  \"source_id\": \"D-101, D-103\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Two concurrent version 1 definitions active: D-101 (finance), D-103 (sales)\"\n}\n```\n", "obj": {"answer": "version 1", "owner_team": null, "source_id": "D-101, D-103", "policy_id": null, "unanswerable": false, "stale": false, "note": "Two concurrent version 1 definitions active: D-101 (finance), D-103 (sales)"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "HIST-03", "family": "HIST", "rep": 1, "latency_s": 82.4, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"an outage breaching the uptime term of a signed contract\",\n  \"owner_team\": \"Engineering\",\n  \"source_id\": \"D-403\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Only D-403 definition was active on 2025-12-01\"\n}\n```\n", "obj": {"answer": "an outage breaching the uptime term of a signed contract", "owner_team": "Engineering", "source_id": "D-403", "policy_id": null, "unanswerable": false, "stale": false, "note": "Only D-403 definition was active on 2025-12-01"}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "PROV-03", "family": "PROV", "rep": 0, "latency_s": 34.41, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"LEG-POL-04\",\n  \"owner_team\": \"legal\",\n  \"source_id\": \"LEG-POL-04\",\n  \"policy_id\": \"D-802\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal policy D-802 (5yr) takes precedence over support wiki D-803 (7yr)\"\n}\n```\n", "obj": {"answer": "LEG-POL-04", "owner_team": "legal", "source_id": "LEG-POL-04", "policy_id": "D-802", "unanswerable": false, "stale": false, "note": "Legal policy D-802 (5yr) takes precedence over support wiki D-803 (7yr)"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dimension", "item": "PROV-03", "family": "PROV", "rep": 1, "latency_s": 18.31, "context_chars": 11191, "raw": "```json\n{\"answer\": \"LEG-POL-04: customer records kept for five years after contract ends\", \"owner_team\": \"legal\", \"source_id\": \"D-802\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Current authoritative definition valid from 2026-06-01\"}\n```\n", "obj": {"answer": "LEG-POL-04: customer records kept for five years after contract ends", "owner_team": "legal", "source_id": "D-802", "policy_id": null, "unanswerable": false, "stale": false, "note": "Current authoritative definition valid from 2026-06-01"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "ROUTE-03", "family": "ROUTE", "rep": 0, "latency_s": 12.01, "context_chars": 2694, "raw": "```json\n{\"answer\": \"legal\", \"owner_team\": \"legal\", \"source_id\": null, \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Legal team owns data retention policy.\"}\n```\n", "obj": {"answer": "legal", "owner_team": "legal", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal team owns data retention policy."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "GAP-02", "family": "GAP", "rep": 0, "latency_s": 11.79, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Memory contains only planned headcount growth (engineering +11 in 2027), not current total.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Memory contains only planned headcount growth (engineering +11 in 2027), not current total."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "GAP-03", "family": "GAP", "rep": 1, "latency_s": 10.58, "context_chars": 6474, "raw": "```json\n{\"answer\": null, \"owner_team\": null, \"source_id\": null, \"policy_id\": null, \"unanswerable\": true, \"stale\": false, \"note\": \"No definition of gross margin found in organizational memory.\"}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "No definition of gross margin found in organizational memory."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 1, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 0, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 1, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "PROV-04", "family": "PROV", "rep": 1, "latency_s": 18.17, "context_chars": 6474, "raw": "```json\n{\"answer\": \"SUP-POL-03\", \"owner_team\": \"support\", \"source_id\": \"D-402\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"D-402 v2 effective 2026-01-15, supersedes earlier definitions\"}\n```\n", "obj": {"answer": "SUP-POL-03", "owner_team": "support", "source_id": "D-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "D-402 v2 effective 2026-01-15, supersedes earlier definitions"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "CONF-02", "family": "CONF", "rep": 0, "latency_s": 16.79, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": null,\n  \"owner_team\": null,\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": true,\n  \"stale\": false,\n  \"note\": \"Sales and Support use different definitions of active_customer; no resolution policy provided.\"\n}\n```\n", "obj": {"answer": null, "owner_team": null, "source_id": null, "policy_id": null, "unanswerable": true, "stale": false, "note": "Sales and Support use different definitions of active_customer; no resolution policy provided."}, "parsed": 1, "answer_correct": 0, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 1, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "AUTH-03", "family": "AUTH", "rep": 1, "latency_s": 21.19, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"318\",\n  \"owner_team\": \"sales\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales count based on live contracts; Support reports 402 using stricter criteria\"\n}\n```\n", "obj": {"answer": "318", "owner_team": "sales", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales count based on live contracts; Support reports 402 using stricter criteria"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "CONF-06", "family": "CONF", "rep": 0, "latency_s": 10.97, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"People's definition applies: a person engaged under a services agreement with no employment relationship.\",\n  \"owner_team\": \"People\",\n  \"source_id\": \"PPL-POL-02\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"People owns the definition; their record prevails over Finance's per CP-ORG.\"\n}\n```\n", "obj": {"answer": "People's definition applies: a person engaged under a services agreement with no employment relationship.", "owner_team": "People", "source_id": "PPL-POL-02", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "People owns the definition; their record prevails over Finance's per CP-ORG."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 1, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 1, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "STALE-04", "family": "STALE", "rep": 1, "latency_s": 9.56, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"Yes. A qualified lead is defined as 'a lead with a named budget holder and a stated timeframe' (valid from 2025-10-01 with no end date).\",\n  \"owner_team\": \"sales\",\n  \"source_id\": \"D-701\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Definition remains in force with open validity.\"\n}\n```\n", "obj": {"answer": "Yes. A qualified lead is defined as 'a lead with a named budget holder and a stated timeframe' (valid from 2025-10-01 with no end date).", "owner_team": "sales", "source_id": "D-701", "policy_id": null, "unanswerable": false, "stale": false, "note": "Definition remains in force with open validity."}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 1, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "dim_no_gov", "item": "HIST-03", "family": "HIST", "rep": 1, "latency_s": 35.28, "context_chars": 6474, "raw": "```json\n{\"answer\": \"1\", \"owner_team\": \"engineering\", \"source_id\": \"D-403\", \"policy_id\": null, \"unanswerable\": false, \"stale\": true, \"note\": \"Two conflicting version 1 definitions both applied; selected more recent\"}\n```\n", "obj": {"answer": "1", "owner_team": "engineering", "source_id": "D-403", "policy_id": null, "unanswerable": false, "stale": true, "note": "Two conflicting version 1 definitions both applied; selected more recent"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "AUTH-01", "family": "AUTH", "rep": 0, "latency_s": 15.32, "context_chars": 6474, "raw": "```json\n{\"answer\": \"9.8m EUR from finance\", \"owner_team\": \"finance\", \"source_id\": \"N-102\", \"policy_id\": null, \"unanswerable\": false, \"stale\": false, \"note\": \"Finance figure is current; sales figure expired 2026-04-30.\"}\n```\n", "obj": {"answer": "9.8m EUR from finance", "owner_team": "finance", "source_id": "N-102", "policy_id": null, "unanswerable": false, "stale": false, "note": "Finance figure is current; sales figure expired 2026-04-30."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "AUTH-01", "family": "AUTH", "rep": 1, "latency_s": 21.08, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"9.8m EUR\",\n  \"owner_team\": \"finance\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": true,\n  \"note\": \"Finance ARR definition refined 2026-04-01; latest data point is 2026-01-01.\"\n}\n```\n", "obj": {"answer": "9.8m EUR", "owner_team": "finance", "source_id": null, "policy_id": null, "unanswerable": false, "stale": true, "note": "Finance ARR definition refined 2026-04-01; latest data point is 2026-01-01."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "dim_no_gov", "item": "STALE-02", "family": "STALE", "rep": 1, "latency_s": 40.55, "context_chars": 6474, "raw": "```json\n{\n  \"answer\": \"No; conflicts with Support's definition D-402, both active.\",\n  \"owner_team\": \"engineering\",\n  \"source_id\": \"D-403\",\n  \"policy_id\": \"D-402\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Two conflicting incident_severity_1 definitions active.\"\n}\n```\n", "obj": {"answer": "No; conflicts with Support's definition D-402, both active.", "owner_team": "engineering", "source_id": "D-403", "policy_id": "D-402", "unanswerable": false, "stale": false, "note": "Two conflicting incident_severity_1 definitions active."}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 1}
{"arm": "wiki", "item": "PROV-02", "family": "PROV", "rep": 1, "latency_s": 16.07, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Sales in document SAL-DECK-Q1\",\n  \"owner_team\": \"Finance\",\n  \"source_id\": \"SAL-DECK-Q1\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Sales reported 12.4m EUR for Q1 2026 (2026-01-01 to 2026-04-30)\"\n}\n```\n", "obj": {"answer": "Sales in document SAL-DECK-Q1", "owner_team": "Finance", "source_id": "SAL-DECK-Q1", "policy_id": null, "unanswerable": false, "stale": false, "note": "Sales reported 12.4m EUR for Q1 2026 (2026-01-01 to 2026-04-30)"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "graph_catalog", "item": "HIST-04", "family": "HIST", "rep": 0, "latency_s": 17.32, "context_chars": 10621, "raw": "```json\n{\n  \"answer\": \"an outage affecting any customer with no workaround, or any confirmed data loss\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": \"CP-ORG\",\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"D-402 from Support prevailed over competing D-403 from Engineering per CP-ORG rule.\"\n}\n```\n", "obj": {"answer": "an outage affecting any customer with no workaround, or any confirmed data loss", "owner_team": "Support", "source_id": "D-402", "policy_id": "CP-ORG", "unanswerable": false, "stale": false, "note": "D-402 from Support prevailed over competing D-403 from Engineering per CP-ORG rule."}, "parsed": 1, "answer_correct": 0, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "wiki", "item": "AUTH-04", "family": "AUTH", "rep": 0, "latency_s": 18.95, "context_chars": 6618, "raw": "```json\n{\n  \"answer\": \"Legal (specifically the General Counsel)\",\n  \"owner_team\": \"Legal\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Legal owns the customer data retention period definition\"\n}\n```\n", "obj": {"answer": "Legal (specifically the General Counsel)", "owner_team": "Legal", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Legal owns the customer data retention period definition"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}
{"arm": "bitemporal", "item": "PROV-04", "family": "PROV", "rep": 1, "latency_s": 34.2, "context_chars": 8932, "raw": "```json\n{\n  \"answer\": \"SUP-POL-03\",\n  \"owner_team\": \"Support\",\n  \"source_id\": \"D-402\",\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Conflict: D-403 also defines severity 1\"\n}\n```\n", "obj": {"answer": "SUP-POL-03", "owner_team": "Support", "source_id": "D-402", "policy_id": null, "unanswerable": false, "stale": false, "note": "Conflict: D-403 also defines severity 1"}, "parsed": 1, "answer_correct": 1, "owner_correct": 0, "policy_correct": 0, "source_correct": 1, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 0, "scored_policy": 0, "scored_source": 1, "scored_abst": 0, "scored_stale": 0}
{"arm": "vector_memory", "item": "AUTH-02", "family": "AUTH", "rep": 0, "latency_s": 37.32, "context_chars": 2694, "raw": "```json\n{\n  \"answer\": \"Sales (2024-09-01) and Support (2025-02-01) define it differently. Sales: 'customer with live contract on reporting date'; Support: 'customer opened ticket in last 90 days'\",\n  \"owner_team\": \"sales, support\",\n  \"source_id\": null,\n  \"policy_id\": null,\n  \"unanswerable\": false,\n  \"stale\": false,\n  \"note\": \"Two teams own conflicting definitions without stated precedence\"\n}\n```\n", "obj": {"answer": "Sales (2024-09-01) and Support (2025-02-01) define it differently. Sales: 'customer with live contract on reporting date'; Support: 'customer opened ticket in last 90 days'", "owner_team": "sales, support", "source_id": null, "policy_id": null, "unanswerable": false, "stale": false, "note": "Two teams own conflicting definitions without stated precedence"}, "parsed": 1, "answer_correct": 1, "owner_correct": 1, "policy_correct": 0, "source_correct": 0, "abstained": 0, "false_abstention": 0, "stale_correct": 0, "overclaim": 0, "fabricated": 0, "scored_answer": 1, "scored_owner": 1, "scored_policy": 0, "scored_source": 0, "scored_abst": 0, "scored_stale": 0}