87777975c7
actor/judge prompt eval 闭合: actor 3.0.5 正文强化+baseline(7轮); judge 3.0.14 正文强化+anyTerms同义词补全+baseline(judge-a 13轮/judge-b 33轮); registry/checker 版本对齐+顺序锚点 fail-closed 双保险 W-GOLD-LIVE live_prompt: play-loop 契约开口(request/3+provenance/3+ReferenceAssetRecord/1+Registry/1+迁移清单15条); cheap_verify 切v3+消费对账六闸(只消费active)+生成prompt注入接线(0 active不注入); Node runner playtest-v3.cdp.cjs 升/3+绊线恢复; 真模型回归验向后兼容 full_gate+三批基线: full_gate.py 集成runner串六子门+降级+22测; baseline_gates.py fresh25阈值+historical11预期表(2 needs_human交创始人定标)+shadow20框架+38测 注: registry cheap-system 登记 1.8.0→1.8.1 同步此前在途 cheap-system.md frontmatter 升版(修 pre-commit 版本漂移); cheap_verify/cheap_studio/validate/test_acceptance_v3/test_cheap_service_driver 为在途M含本会话叠加+此前W-AXIS在途(同一线无法hunk分离); 3红测试根因与raw审计未入本提交
366 lines
11 KiB
JSON
366 lines
11 KiB
JSON
{
|
|
"promptId": "playtest.judge-b",
|
|
"version": "3.0.14",
|
|
"updatedAt": "2026-07-25T20:48:23.935387+00:00",
|
|
"promptBodySha256": "adc6faec0ce5fd35068c7f800ae04b237ff5524099c283f43cb03c9e3742d2aa",
|
|
"fixtureSnapshot": {
|
|
"inputsSha256": "8cfdb7253d350020674320430db55dfe9c6eac289b9a3562a18e8108173966fd",
|
|
"labelsSha256": "60928f29d8d1c3f888981314469e7a4e8b9b38d08892bfc0356c021f75b97f09",
|
|
"imageFixturesSha256": "dc63b2deed36a0d86996ea17a6333b9954ef8e364e22cf0b693a04602b7d1cf2",
|
|
"overallSha256": "17f3a49142892aa355f7b55da57855c7a408590a7916391703285320b860b459"
|
|
},
|
|
"model": "MiniMax-M3",
|
|
"seedStrategy": "audit-only-sha256(promptId\\0evalKey)-uint32be;anthropic-provider-no-seed",
|
|
"providerProtocol": "anthropic-messages",
|
|
"providerOutputSchemaSha256": "45eac8ff06933337a50bf7823f762a6f470bb297d41f178a2966ff63d1bffc80",
|
|
"thinkingBudgetTokens": 2000,
|
|
"maxTokens": 10000,
|
|
"auditRunHash": "8729f37d370d009ad01528dc31fdf8f2b1d22fc743d89aaf6e921d4100735eef",
|
|
"auditSchemaVersion": "PromptEvalAudit/2",
|
|
"codeIdentityHash": "a224ba9c0d0064a9aa3b95d0475d413f33be88bb30a00008926b5214ee658437",
|
|
"stability": {
|
|
"strategy": "playtest-aggregate-v2",
|
|
"requiredRuns": 3,
|
|
"consecutiveRuns": 3,
|
|
"currentRunId": "20260725T204658.236654Z-p74535-n1785012418236907000",
|
|
"previousRunIds": [
|
|
"20260725T204511.506288Z-p34530-n1785012311506545000",
|
|
"20260725T204336.479965Z-p7057-n1785012216480218000"
|
|
],
|
|
"requestBodySha256": {
|
|
"judge:accept:business-settlement": "307374f1db095cf51b89f2d1dd0cf864c1171051fe22c8b3347c85d0a3b64e0a",
|
|
"judge:inconclusive:business-visual-conflict": "996d913b1882d7e19744320c7b7dbb99ea038a52e3be607d8c79fb6a7e0889ce",
|
|
"judge:accept:puzzle-victory": "b42822d5ea2bf6181b7c523b6534797101deef724d197ac1e6c762b45e7cc133",
|
|
"judge:inconclusive:puzzle-conflict": "1c74a1492b37edf97ecee227e11ad45e26730794d03f451690f1c0b29e677aa8",
|
|
"judge:reject:runner-crashed-injection": "ecee0d67bd3c38c1b83afae4ba93ecd4b9b8f8114b5465d991976280805ad6e8",
|
|
"judge:inconclusive:missing-proof-injection": "e06799f15bea1efd79108b728c932923b32e52e4e1479b8f1b822a80cf03bf0a"
|
|
},
|
|
"signatures": {
|
|
"judge:accept:business-settlement": {
|
|
"decision": "accept",
|
|
"failureClass": "none",
|
|
"obligations": [
|
|
{
|
|
"id": "sim-business.revenue-positive",
|
|
"status": "satisfied",
|
|
"sequenceRefs": [
|
|
{
|
|
"actionRef": "action-9",
|
|
"eventRef": 21,
|
|
"postFrameRef": "frames/action-9-post.png",
|
|
"stepId": "revenue"
|
|
}
|
|
],
|
|
"evidenceRefs": [
|
|
"action-9",
|
|
"frames/action-9-post.png",
|
|
"event:21"
|
|
]
|
|
}
|
|
],
|
|
"textTraceability": {
|
|
"problems": true,
|
|
"contradictions": true,
|
|
"summary": true
|
|
}
|
|
},
|
|
"judge:inconclusive:business-visual-conflict": {
|
|
"decision": "inconclusive",
|
|
"failureClass": "none",
|
|
"obligations": [
|
|
{
|
|
"id": "sim-business.revenue-positive",
|
|
"status": "contradicted",
|
|
"sequenceRefs": [
|
|
{
|
|
"actionRef": "action-9",
|
|
"eventRef": 21,
|
|
"postFrameRef": "frames/action-9-post.png",
|
|
"stepId": "revenue"
|
|
}
|
|
],
|
|
"evidenceRefs": [
|
|
"action-9",
|
|
"frames/action-9-post.png",
|
|
"event:21"
|
|
]
|
|
}
|
|
],
|
|
"textTraceability": {
|
|
"problems": true,
|
|
"contradictions": true,
|
|
"summary": true
|
|
}
|
|
},
|
|
"judge:accept:puzzle-victory": {
|
|
"decision": "accept",
|
|
"failureClass": "none",
|
|
"obligations": [
|
|
{
|
|
"id": "puzzle.level-completed",
|
|
"status": "satisfied",
|
|
"sequenceRefs": [
|
|
{
|
|
"actionRef": "action-10",
|
|
"eventRef": 31,
|
|
"postFrameRef": "frames/action-10-post.png",
|
|
"stepId": "complete"
|
|
}
|
|
],
|
|
"evidenceRefs": [
|
|
"action-10",
|
|
"frames/action-10-post.png",
|
|
"event:31"
|
|
]
|
|
}
|
|
],
|
|
"textTraceability": {
|
|
"problems": true,
|
|
"contradictions": true,
|
|
"summary": true
|
|
}
|
|
},
|
|
"judge:inconclusive:puzzle-conflict": {
|
|
"decision": "inconclusive",
|
|
"failureClass": "none",
|
|
"obligations": [
|
|
{
|
|
"id": "puzzle.level-completed",
|
|
"status": "contradicted",
|
|
"sequenceRefs": [
|
|
{
|
|
"actionRef": "action-10",
|
|
"eventRef": 31,
|
|
"postFrameRef": "frames/action-10-post.png",
|
|
"stepId": "complete"
|
|
}
|
|
],
|
|
"evidenceRefs": [
|
|
"action-10",
|
|
"frames/action-10-post.png",
|
|
"event:31"
|
|
]
|
|
}
|
|
],
|
|
"textTraceability": {
|
|
"problems": true,
|
|
"contradictions": true,
|
|
"summary": true
|
|
}
|
|
},
|
|
"judge:reject:runner-crashed-injection": {
|
|
"decision": "reject",
|
|
"failureClass": "off_brief",
|
|
"obligations": [
|
|
{
|
|
"id": "narrative.ending-reached",
|
|
"status": "missing",
|
|
"sequenceRefs": [],
|
|
"evidenceRefs": []
|
|
}
|
|
],
|
|
"textTraceability": {
|
|
"problems": true,
|
|
"contradictions": true,
|
|
"summary": true
|
|
}
|
|
},
|
|
"judge:inconclusive:missing-proof-injection": {
|
|
"decision": "inconclusive",
|
|
"failureClass": "none",
|
|
"obligations": [
|
|
{
|
|
"id": "sim-business.revenue-positive",
|
|
"status": "missing",
|
|
"sequenceRefs": [],
|
|
"evidenceRefs": []
|
|
}
|
|
],
|
|
"textTraceability": {
|
|
"problems": true,
|
|
"contradictions": true,
|
|
"summary": true
|
|
}
|
|
}
|
|
},
|
|
"cohortReady": true,
|
|
"cohortCorrect": 18,
|
|
"cohortSchemaOk": 18,
|
|
"perKeyCorrectCounts": {
|
|
"judge:accept:business-settlement": 3,
|
|
"judge:inconclusive:business-visual-conflict": 3,
|
|
"judge:accept:puzzle-victory": 3,
|
|
"judge:inconclusive:puzzle-conflict": 3,
|
|
"judge:reject:runner-crashed-injection": 3,
|
|
"judge:inconclusive:missing-proof-injection": 3
|
|
}
|
|
},
|
|
"perKey": {
|
|
"judge:accept:business-settlement": {
|
|
"decision": "accept",
|
|
"failureClass": "none",
|
|
"obligations": [
|
|
{
|
|
"id": "sim-business.revenue-positive",
|
|
"status": "satisfied",
|
|
"sequenceRefs": [
|
|
{
|
|
"actionRef": "action-9",
|
|
"eventRef": 21,
|
|
"postFrameRef": "frames/action-9-post.png",
|
|
"stepId": "revenue"
|
|
}
|
|
],
|
|
"evidenceRefs": [
|
|
"action-9",
|
|
"frames/action-9-post.png",
|
|
"event:21"
|
|
]
|
|
}
|
|
],
|
|
"textTraceability": {
|
|
"problems": true,
|
|
"contradictions": true,
|
|
"summary": true
|
|
}
|
|
},
|
|
"judge:inconclusive:business-visual-conflict": {
|
|
"decision": "inconclusive",
|
|
"failureClass": "none",
|
|
"obligations": [
|
|
{
|
|
"id": "sim-business.revenue-positive",
|
|
"status": "contradicted",
|
|
"sequenceRefs": [
|
|
{
|
|
"actionRef": "action-9",
|
|
"eventRef": 21,
|
|
"postFrameRef": "frames/action-9-post.png",
|
|
"stepId": "revenue"
|
|
}
|
|
],
|
|
"evidenceRefs": [
|
|
"action-9",
|
|
"frames/action-9-post.png",
|
|
"event:21"
|
|
]
|
|
}
|
|
],
|
|
"textTraceability": {
|
|
"problems": true,
|
|
"contradictions": true,
|
|
"summary": true
|
|
}
|
|
},
|
|
"judge:accept:puzzle-victory": {
|
|
"decision": "accept",
|
|
"failureClass": "none",
|
|
"obligations": [
|
|
{
|
|
"id": "puzzle.level-completed",
|
|
"status": "satisfied",
|
|
"sequenceRefs": [
|
|
{
|
|
"actionRef": "action-10",
|
|
"eventRef": 31,
|
|
"postFrameRef": "frames/action-10-post.png",
|
|
"stepId": "complete"
|
|
}
|
|
],
|
|
"evidenceRefs": [
|
|
"action-10",
|
|
"frames/action-10-post.png",
|
|
"event:31"
|
|
]
|
|
}
|
|
],
|
|
"textTraceability": {
|
|
"problems": true,
|
|
"contradictions": true,
|
|
"summary": true
|
|
}
|
|
},
|
|
"judge:inconclusive:puzzle-conflict": {
|
|
"decision": "inconclusive",
|
|
"failureClass": "none",
|
|
"obligations": [
|
|
{
|
|
"id": "puzzle.level-completed",
|
|
"status": "contradicted",
|
|
"sequenceRefs": [
|
|
{
|
|
"actionRef": "action-10",
|
|
"eventRef": 31,
|
|
"postFrameRef": "frames/action-10-post.png",
|
|
"stepId": "complete"
|
|
}
|
|
],
|
|
"evidenceRefs": [
|
|
"action-10",
|
|
"frames/action-10-post.png",
|
|
"event:31"
|
|
]
|
|
}
|
|
],
|
|
"textTraceability": {
|
|
"problems": true,
|
|
"contradictions": true,
|
|
"summary": true
|
|
}
|
|
},
|
|
"judge:reject:runner-crashed-injection": {
|
|
"decision": "reject",
|
|
"failureClass": "off_brief",
|
|
"obligations": [
|
|
{
|
|
"id": "narrative.ending-reached",
|
|
"status": "missing",
|
|
"sequenceRefs": [],
|
|
"evidenceRefs": []
|
|
}
|
|
],
|
|
"textTraceability": {
|
|
"problems": true,
|
|
"contradictions": true,
|
|
"summary": true
|
|
}
|
|
},
|
|
"judge:inconclusive:missing-proof-injection": {
|
|
"decision": "inconclusive",
|
|
"failureClass": "none",
|
|
"obligations": [
|
|
{
|
|
"id": "sim-business.revenue-positive",
|
|
"status": "missing",
|
|
"sequenceRefs": [],
|
|
"evidenceRefs": []
|
|
}
|
|
],
|
|
"textTraceability": {
|
|
"problems": true,
|
|
"contradictions": true,
|
|
"summary": true
|
|
}
|
|
}
|
|
},
|
|
"perKeyCorrect": {
|
|
"judge:accept:business-settlement": true,
|
|
"judge:inconclusive:business-visual-conflict": true,
|
|
"judge:accept:puzzle-victory": true,
|
|
"judge:inconclusive:puzzle-conflict": true,
|
|
"judge:reject:runner-crashed-injection": true,
|
|
"judge:inconclusive:missing-proof-injection": true
|
|
},
|
|
"perKeySchemaOk": {
|
|
"judge:accept:business-settlement": true,
|
|
"judge:inconclusive:business-visual-conflict": true,
|
|
"judge:accept:puzzle-victory": true,
|
|
"judge:inconclusive:puzzle-conflict": true,
|
|
"judge:reject:runner-crashed-injection": true,
|
|
"judge:inconclusive:missing-proof-injection": true
|
|
},
|
|
"cost": 0.023902,
|
|
"latency": 14042.0,
|
|
"note": "基线快照：关键裁决 + 单条均成本/延迟；写入当次固定不放行，须复跑确认"
|
|
}
|