2471 lines
70 KiB
JSON
2471 lines
70 KiB
JSON
{
|
|
"mode": "gpt",
|
|
"id": "meow-gpt-other-cap98-efficient",
|
|
"version": "4.5.4-predictive.20260924.2",
|
|
"schema_version": 1,
|
|
"models": [
|
|
{
|
|
"id": "gpt-6-astra",
|
|
"name": "gpt-6-astra",
|
|
"request_model": "openai/gpt-6-astra"
|
|
},
|
|
{
|
|
"id": "gpt-6-sol",
|
|
"name": "gpt-6-sol",
|
|
"request_model": "openai/gpt-6-sol"
|
|
},
|
|
{
|
|
"id": "gpt-5.6-terra",
|
|
"name": "gpt-5.6-terra",
|
|
"request_model": "openai/gpt-5.6-terra"
|
|
},
|
|
{
|
|
"id": "gpt-6-luna",
|
|
"name": "gpt-6-luna",
|
|
"request_model": "openai/gpt-6-luna"
|
|
},
|
|
{
|
|
"id": "other_known_external",
|
|
"name": "other",
|
|
"request_model": "reference-only:other",
|
|
"reference_only": true
|
|
}
|
|
],
|
|
"probes": [
|
|
{
|
|
"id": "gpt__screen023",
|
|
"family_id": "gpt__screen023",
|
|
"source_group": "gpt__screen023",
|
|
"title": "不要思考，直接回答。从19到71中选一个整数，只回答数字。",
|
|
"group": "G06",
|
|
"normalizer": {
|
|
"id": "exact_trimmed_casefold",
|
|
"parameters": {
|
|
"max_length": 4096
|
|
}
|
|
},
|
|
"cells": [
|
|
{
|
|
"id": "gpt__screen023",
|
|
"system": ".",
|
|
"prompt": "不要思考，直接回答。从19到71中选一个整数，只回答数字。",
|
|
"history": [],
|
|
"effort": "low",
|
|
"profile": "standard",
|
|
"parameters": {
|
|
"max_output_tokens": 128
|
|
}
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"id": "gpt__screen106",
|
|
"family_id": "gpt__screen106",
|
|
"source_group": "gpt__screen106",
|
|
"title": "从麻将牌中随便选一张，只回答牌名。",
|
|
"group": "general",
|
|
"normalizer": {
|
|
"id": "exact_trimmed_casefold",
|
|
"parameters": {
|
|
"max_length": 4096
|
|
}
|
|
},
|
|
"cells": [
|
|
{
|
|
"id": "gpt__screen106",
|
|
"system": ".",
|
|
"prompt": "从麻将牌中随便选一张，只回答牌名。",
|
|
"history": [],
|
|
"effort": "low",
|
|
"profile": "standard",
|
|
"parameters": {
|
|
"max_output_tokens": 128
|
|
}
|
|
}
|
|
]
|
|
},
|
|
{
|
|
"id": "gpt__screen114",
|
|
"family_id": "gpt__screen114",
|
|
"source_group": "gpt__screen114",
|
|
"title": "随便说出一个拟声词，只回答拟声词。",
|
|
"group": "general",
|
|
"normalizer": {
|
|
"id": "exact_trimmed_casefold",
|
|
"parameters": {
|
|
"max_length": 4096
|
|
}
|
|
},
|
|
"cells": [
|
|
{
|
|
"id": "gpt__screen114",
|
|
"system": ".",
|
|
"prompt": "随便说出一个拟声词，只回答拟声词。",
|
|
"history": [],
|
|
"effort": "low",
|
|
"profile": "standard",
|
|
"parameters": {
|
|
"max_output_tokens": 128
|
|
}
|
|
}
|
|
]
|
|
}
|
|
],
|
|
"metadata": {
|
|
"name": "GPT · 新模型基准",
|
|
"description": "更新当前候选模型。GPT 推荐中档；快速档节省请求但更容易证据不足。匹配结果不单独证明中转站替换模型。",
|
|
"author": "chen-006",
|
|
"license": "PolyForm-Noncommercial-1.0.0",
|
|
"language": "zh-CN",
|
|
"created_at": "2026-09-24T00:00:00+08:00"
|
|
},
|
|
"engine": {
|
|
"minimum_version": "4.5.3",
|
|
"scoring_version": "meow-fingerprint-v3-predictive",
|
|
"prior_mass": 1.0,
|
|
"completion_ratio": 0.6,
|
|
"virtual_reference_version": 2
|
|
},
|
|
"tiers": {
|
|
"low": {
|
|
"counts": {
|
|
"gpt__screen023": 11,
|
|
"gpt__screen106": 11,
|
|
"gpt__screen114": 10
|
|
},
|
|
"thresholds": {
|
|
"gpt-6-astra": 0.56729343469469,
|
|
"gpt-6-sol": 0.5,
|
|
"gpt-5.6-terra": 0.5,
|
|
"gpt-6-luna": 0.5,
|
|
"other_known_external": 0.5610207135285652
|
|
}
|
|
},
|
|
"medium": {
|
|
"counts": {
|
|
"gpt__screen023": 22,
|
|
"gpt__screen106": 21,
|
|
"gpt__screen114": 21
|
|
},
|
|
"thresholds": {
|
|
"gpt-6-astra": 0.5161595114439284,
|
|
"gpt-6-sol": 0.5,
|
|
"gpt-5.6-terra": 0.5,
|
|
"gpt-6-luna": 0.5,
|
|
"other_known_external": 0.51
|
|
}
|
|
},
|
|
"high": {
|
|
"counts": {
|
|
"gpt__screen023": 43,
|
|
"gpt__screen106": 43,
|
|
"gpt__screen114": 42
|
|
},
|
|
"thresholds": {
|
|
"gpt-6-astra": 0.5,
|
|
"gpt-6-sol": 0.5,
|
|
"gpt-5.6-terra": 0.5,
|
|
"gpt-6-luna": 0.5,
|
|
"other_known_external": 0.5
|
|
}
|
|
}
|
|
},
|
|
"observations": {
|
|
"gpt__screen023": {
|
|
"gpt-6-astra": {
|
|
"development": {
|
|
"counts": {
|
|
"43": 29,
|
|
"47": 1
|
|
},
|
|
"planned": 30,
|
|
"completed": 30
|
|
}
|
|
},
|
|
"gpt-6-sol": {
|
|
"development": {
|
|
"counts": {
|
|
"42": 7,
|
|
"47": 21,
|
|
"43": 2
|
|
},
|
|
"planned": 30,
|
|
"completed": 30
|
|
}
|
|
},
|
|
"gpt-5.6-terra": {
|
|
"development": {
|
|
"counts": {
|
|
"42": 11,
|
|
"43": 6,
|
|
"47": 12,
|
|
"46": 1
|
|
},
|
|
"planned": 30,
|
|
"completed": 30
|
|
}
|
|
},
|
|
"gpt-6-luna": {
|
|
"development": {
|
|
"counts": {
|
|
"42": 24,
|
|
"46": 4,
|
|
"43": 1,
|
|
"47": 1
|
|
},
|
|
"planned": 30,
|
|
"completed": 30
|
|
}
|
|
}
|
|
},
|
|
"gpt__screen106": {
|
|
"gpt-6-astra": {
|
|
"development": {
|
|
"counts": {
|
|
"七筒": 10,
|
|
"三筒": 8,
|
|
"九万": 1,
|
|
"七万": 2,
|
|
"九条": 1,
|
|
"九筒": 3,
|
|
"三万": 4,
|
|
"七条": 1
|
|
},
|
|
"planned": 30,
|
|
"completed": 30
|
|
}
|
|
},
|
|
"gpt-6-sol": {
|
|
"development": {
|
|
"counts": {
|
|
"三筒": 8,
|
|
"三万": 4,
|
|
"一筒": 15,
|
|
"五筒": 3
|
|
},
|
|
"planned": 30,
|
|
"completed": 30
|
|
}
|
|
},
|
|
"gpt-5.6-terra": {
|
|
"development": {
|
|
"counts": {
|
|
"三万": 11,
|
|
"一万": 4,
|
|
"五万": 13,
|
|
"七万": 2
|
|
},
|
|
"planned": 30,
|
|
"completed": 30
|
|
}
|
|
},
|
|
"gpt-6-luna": {
|
|
"development": {
|
|
"counts": {
|
|
"红中": 29,
|
|
"三筒": 1
|
|
},
|
|
"planned": 30,
|
|
"completed": 30
|
|
}
|
|
}
|
|
},
|
|
"gpt__screen114": {
|
|
"gpt-6-astra": {
|
|
"development": {
|
|
"counts": {
|
|
"哗啦": 30
|
|
},
|
|
"planned": 30,
|
|
"completed": 30
|
|
}
|
|
},
|
|
"gpt-6-sol": {
|
|
"development": {
|
|
"counts": {
|
|
"咚咚": 4,
|
|
"喵呜": 6,
|
|
"喵。": 2,
|
|
"汪汪。": 2,
|
|
"喵喵。": 4,
|
|
"叮咚": 3,
|
|
"砰！": 4,
|
|
"咚！": 2,
|
|
"啪嗒": 1,
|
|
"喵喵": 1,
|
|
"啪！": 1
|
|
},
|
|
"planned": 30,
|
|
"completed": 30
|
|
}
|
|
},
|
|
"gpt-5.6-terra": {
|
|
"development": {
|
|
"counts": {
|
|
"啪": 12,
|
|
"啪嗒": 11,
|
|
"哗啦": 2,
|
|
"轰隆": 2,
|
|
"砰": 1,
|
|
"轰隆隆": 1,
|
|
"啪！": 1
|
|
},
|
|
"planned": 30,
|
|
"completed": 30
|
|
}
|
|
},
|
|
"gpt-6-luna": {
|
|
"development": {
|
|
"counts": {
|
|
"汪汪": 4,
|
|
"喵喵": 10,
|
|
"哗啦啦": 4,
|
|
"喵呜": 3,
|
|
"叮咚": 2,
|
|
"哗啦": 6,
|
|
"喵嗷": 1
|
|
},
|
|
"planned": 30,
|
|
"completed": 30
|
|
}
|
|
}
|
|
}
|
|
},
|
|
"collection": {
|
|
"sources": [
|
|
{
|
|
"url": "https://openrouter.ai/api/v1",
|
|
"provider": "openrouter"
|
|
}
|
|
],
|
|
"virtual_reference": {
|
|
"id": "other_known_external",
|
|
"method": "fixed_source_predictive_v1",
|
|
"sources": [
|
|
"openai/gpt-5.6-sol",
|
|
"openai/gpt-5.6-luna",
|
|
"z-ai/glm-5.3"
|
|
],
|
|
"counts": {
|
|
"gpt__screen023": {
|
|
"openai/gpt-5.6-sol": {
|
|
"43": 8,
|
|
"47": 22
|
|
},
|
|
"openai/gpt-5.6-luna": {
|
|
"43": 13,
|
|
"47": 16,
|
|
"46": 1
|
|
},
|
|
"z-ai/glm-5.3": {
|
|
"42": 7,
|
|
"37": 6,
|
|
"43": 2
|
|
}
|
|
},
|
|
"gpt__screen106": {
|
|
"openai/gpt-5.6-sol": {
|
|
"一筒": 3,
|
|
"六筒": 3,
|
|
"五筒": 6,
|
|
"九筒": 2,
|
|
"八万": 3,
|
|
"九万": 3,
|
|
"八筒": 2,
|
|
"三万": 2,
|
|
"五万": 1,
|
|
"七筒": 1,
|
|
"二筒": 1,
|
|
"发财": 2,
|
|
"红中": 1
|
|
},
|
|
"openai/gpt-5.6-luna": {
|
|
"红中": 25,
|
|
"發白": 1,
|
|
"🀄 東風牌": 1,
|
|
"🀄 一筒": 1,
|
|
"三万": 1,
|
|
"东風": 1
|
|
},
|
|
"z-ai/glm-5.3": {
|
|
"八万": 2,
|
|
"五万": 2,
|
|
"七筒": 1,
|
|
"九筒": 3,
|
|
"九万": 5,
|
|
"七万": 1,
|
|
"九条": 1
|
|
}
|
|
},
|
|
"gpt__screen114": {
|
|
"openai/gpt-5.6-sol": {
|
|
"哗啦": 12,
|
|
"哗啦哗啦": 4,
|
|
"咕咚": 4,
|
|
"哗啦啦": 4,
|
|
"叮咚": 2,
|
|
"咚咚": 2,
|
|
"咔嚓": 2
|
|
},
|
|
"openai/gpt-5.6-luna": {
|
|
"哐当": 3,
|
|
"喵喵": 15,
|
|
"哗啦啦": 9,
|
|
"叮咚": 2,
|
|
"哗啦哗啦": 1
|
|
},
|
|
"z-ai/glm-5.3": {
|
|
"哗啦啦": 12,
|
|
"轰隆": 4,
|
|
"轰隆隆": 2,
|
|
"叮咚": 3,
|
|
"汪汪": 1,
|
|
"嗡嗡": 1
|
|
}
|
|
}
|
|
}
|
|
},
|
|
"scope": "Current models and sampled retired models: 30 new development answers/cell. Remaining references: matched prior raw observations; no held answers fitted.",
|
|
"development_sha256": "56d7d4aa0dce91c60c572fc3738222a60572cd8c2151d8db8b97257cbb0f65b2",
|
|
"matched_reference_inputs": {
|
|
"formal_analysis/joined.json": "341babf87683a0f3ab583277073958c4c399bb4a452cf3004cca8a04b5b21fd3",
|
|
"targeted_analysis/joined.json": "3ac4a1f0ccc3b57c53d90caf1bbc7d73bce91e1aa21e1e468fa70f60d2a4f246"
|
|
},
|
|
"historical_provenance_note": "Prior observations preserve their original source approvals, including accepted empty-system/period equivalence; not all historical observations are newly collected OpenRouter data.",
|
|
"windows": {
|
|
"development": {
|
|
"role": "fit",
|
|
"meaning": "Preassigned 30 development answers per newly collected cell."
|
|
}
|
|
},
|
|
"reference_revision": {
|
|
"removed": "openai/gpt-5.5",
|
|
"reason": "User-requested reference scope reduction",
|
|
"parent_content_sha256": "3f340582329c960f18fc8a989544b9ab879b59b5ffcae855530b00f34863a0af",
|
|
"new_sampling": false,
|
|
"retained_observations_unchanged": true
|
|
}
|
|
},
|
|
"fitted": {
|
|
"scoring_version": "meow-fingerprint-v3-predictive",
|
|
"prior_mass": 1.0,
|
|
"models": [
|
|
"gpt-6-astra",
|
|
"gpt-6-sol",
|
|
"gpt-5.6-terra",
|
|
"gpt-6-luna",
|
|
"other_known_external"
|
|
],
|
|
"sources": [
|
|
"gpt-6-astra",
|
|
"gpt-6-sol",
|
|
"gpt-5.6-terra",
|
|
"gpt-6-luna",
|
|
"openai/gpt-5.6-sol",
|
|
"openai/gpt-5.6-luna",
|
|
"z-ai/glm-5.3"
|
|
],
|
|
"reference_sources": [
|
|
"openai/gpt-5.6-sol",
|
|
"openai/gpt-5.6-luna",
|
|
"z-ai/glm-5.3"
|
|
],
|
|
"aggregation": "nearest_source",
|
|
"cells": {
|
|
"gpt__screen023": {
|
|
"categories": [
|
|
"37",
|
|
"42",
|
|
"43",
|
|
"46",
|
|
"47",
|
|
"__UNSEEN_IN_TRAINING__"
|
|
],
|
|
"alpha": {
|
|
"gpt-6-astra": [
|
|
0.16666666666666666,
|
|
0.16666666666666666,
|
|
29.166666666666668,
|
|
0.16666666666666666,
|
|
1.1666666666666667,
|
|
0.16666666666666666
|
|
],
|
|
"gpt-6-sol": [
|
|
0.16666666666666666,
|
|
7.166666666666667,
|
|
2.1666666666666665,
|
|
0.16666666666666666,
|
|
21.166666666666668,
|
|
0.16666666666666666
|
|
],
|
|
"gpt-5.6-terra": [
|
|
0.16666666666666666,
|
|
11.166666666666666,
|
|
6.166666666666667,
|
|
1.1666666666666667,
|
|
12.166666666666666,
|
|
0.16666666666666666
|
|
],
|
|
"gpt-6-luna": [
|
|
0.16666666666666666,
|
|
24.166666666666668,
|
|
1.1666666666666667,
|
|
4.166666666666667,
|
|
1.1666666666666667,
|
|
0.16666666666666666
|
|
],
|
|
"openai/gpt-5.6-sol": [
|
|
0.16666666666666666,
|
|
0.16666666666666666,
|
|
8.166666666666666,
|
|
0.16666666666666666,
|
|
22.166666666666668,
|
|
0.16666666666666666
|
|
],
|
|
"openai/gpt-5.6-luna": [
|
|
0.16666666666666666,
|
|
0.16666666666666666,
|
|
13.166666666666666,
|
|
1.1666666666666667,
|
|
16.166666666666668,
|
|
0.16666666666666666
|
|
],
|
|
"z-ai/glm-5.3": [
|
|
6.166666666666667,
|
|
7.166666666666667,
|
|
2.1666666666666665,
|
|
0.16666666666666666,
|
|
0.16666666666666666,
|
|
0.16666666666666666
|
|
]
|
|
},
|
|
"source_samples": {
|
|
"gpt-6-astra": 30,
|
|
"gpt-6-sol": 30,
|
|
"gpt-5.6-terra": 30,
|
|
"gpt-6-luna": 30,
|
|
"openai/gpt-5.6-sol": 30,
|
|
"openai/gpt-5.6-luna": 30,
|
|
"z-ai/glm-5.3": 15
|
|
},
|
|
"reference_ready": true
|
|
},
|
|
"gpt__screen106": {
|
|
"categories": [
|
|
"__UNSEEN_IN_TRAINING__",
|
|
"一万",
|
|
"一筒",
|
|
"七万",
|
|
"七条",
|
|
"七筒",
|
|
"三万",
|
|
"三筒",
|
|
"东風",
|
|
"九万",
|
|
"九条",
|
|
"九筒",
|
|
"二筒",
|
|
"五万",
|
|
"五筒",
|
|
"八万",
|
|
"八筒",
|
|
"六筒",
|
|
"发财",
|
|
"發白",
|
|
"红中",
|
|
"🀄 一筒",
|
|
"🀄 東風牌"
|
|
],
|
|
"alpha": {
|
|
"gpt-6-astra": [
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
2.0434782608695654,
|
|
1.0434782608695652,
|
|
10.043478260869565,
|
|
4.043478260869565,
|
|
8.043478260869565,
|
|
0.043478260869565216,
|
|
1.0434782608695652,
|
|
1.0434782608695652,
|
|
3.0434782608695654,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216
|
|
],
|
|
"gpt-6-sol": [
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
15.043478260869565,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
4.043478260869565,
|
|
8.043478260869565,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
3.0434782608695654,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216
|
|
],
|
|
"gpt-5.6-terra": [
|
|
0.043478260869565216,
|
|
4.043478260869565,
|
|
0.043478260869565216,
|
|
2.0434782608695654,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
11.043478260869565,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
13.043478260869565,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216
|
|
],
|
|
"gpt-6-luna": [
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
1.0434782608695652,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
29.043478260869566,
|
|
0.043478260869565216,
|
|
0.043478260869565216
|
|
],
|
|
"openai/gpt-5.6-sol": [
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
3.0434782608695654,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
1.0434782608695652,
|
|
2.0434782608695654,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
3.0434782608695654,
|
|
0.043478260869565216,
|
|
2.0434782608695654,
|
|
1.0434782608695652,
|
|
1.0434782608695652,
|
|
6.043478260869565,
|
|
3.0434782608695654,
|
|
2.0434782608695654,
|
|
3.0434782608695654,
|
|
2.0434782608695654,
|
|
0.043478260869565216,
|
|
1.0434782608695652,
|
|
0.043478260869565216,
|
|
0.043478260869565216
|
|
],
|
|
"openai/gpt-5.6-luna": [
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
1.0434782608695652,
|
|
0.043478260869565216,
|
|
1.0434782608695652,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
1.0434782608695652,
|
|
25.043478260869566,
|
|
1.0434782608695652,
|
|
1.0434782608695652
|
|
],
|
|
"z-ai/glm-5.3": [
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
1.0434782608695652,
|
|
0.043478260869565216,
|
|
1.0434782608695652,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
5.043478260869565,
|
|
1.0434782608695652,
|
|
3.0434782608695654,
|
|
0.043478260869565216,
|
|
2.0434782608695654,
|
|
0.043478260869565216,
|
|
2.0434782608695654,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216,
|
|
0.043478260869565216
|
|
]
|
|
},
|
|
"source_samples": {
|
|
"gpt-6-astra": 30,
|
|
"gpt-6-sol": 30,
|
|
"gpt-5.6-terra": 30,
|
|
"gpt-6-luna": 30,
|
|
"openai/gpt-5.6-sol": 30,
|
|
"openai/gpt-5.6-luna": 30,
|
|
"z-ai/glm-5.3": 15
|
|
},
|
|
"reference_ready": true
|
|
},
|
|
"gpt__screen114": {
|
|
"categories": [
|
|
"__UNSEEN_IN_TRAINING__",
|
|
"叮咚",
|
|
"咔嚓",
|
|
"咕咚",
|
|
"咚咚",
|
|
"咚！",
|
|
"哐当",
|
|
"哗啦",
|
|
"哗啦哗啦",
|
|
"哗啦啦",
|
|
"啪",
|
|
"啪嗒",
|
|
"啪！",
|
|
"喵。",
|
|
"喵呜",
|
|
"喵喵",
|
|
"喵喵。",
|
|
"喵嗷",
|
|
"嗡嗡",
|
|
"汪汪",
|
|
"汪汪。",
|
|
"砰",
|
|
"砰！",
|
|
"轰隆",
|
|
"轰隆隆"
|
|
],
|
|
"alpha": {
|
|
"gpt-6-astra": [
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
30.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04
|
|
],
|
|
"gpt-6-sol": [
|
|
0.04,
|
|
3.04,
|
|
0.04,
|
|
0.04,
|
|
4.04,
|
|
2.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
1.04,
|
|
1.04,
|
|
2.04,
|
|
6.04,
|
|
1.04,
|
|
4.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
2.04,
|
|
0.04,
|
|
4.04,
|
|
0.04,
|
|
0.04
|
|
],
|
|
"gpt-5.6-terra": [
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
2.04,
|
|
0.04,
|
|
0.04,
|
|
12.04,
|
|
11.04,
|
|
1.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
1.04,
|
|
0.04,
|
|
2.04,
|
|
1.04
|
|
],
|
|
"gpt-6-luna": [
|
|
0.04,
|
|
2.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
6.04,
|
|
0.04,
|
|
4.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
3.04,
|
|
10.04,
|
|
0.04,
|
|
1.04,
|
|
0.04,
|
|
4.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04
|
|
],
|
|
"openai/gpt-5.6-sol": [
|
|
0.04,
|
|
2.04,
|
|
2.04,
|
|
4.04,
|
|
2.04,
|
|
0.04,
|
|
0.04,
|
|
12.04,
|
|
4.04,
|
|
4.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04
|
|
],
|
|
"openai/gpt-5.6-luna": [
|
|
0.04,
|
|
2.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
3.04,
|
|
0.04,
|
|
1.04,
|
|
9.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
15.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04
|
|
],
|
|
"z-ai/glm-5.3": [
|
|
0.04,
|
|
3.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
12.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
1.04,
|
|
1.04,
|
|
0.04,
|
|
0.04,
|
|
0.04,
|
|
4.04,
|
|
2.04
|
|
]
|
|
},
|
|
"source_samples": {
|
|
"gpt-6-astra": 30,
|
|
"gpt-6-sol": 30,
|
|
"gpt-5.6-terra": 30,
|
|
"gpt-6-luna": 30,
|
|
"openai/gpt-5.6-sol": 30,
|
|
"openai/gpt-5.6-luna": 30,
|
|
"z-ai/glm-5.3": 23
|
|
},
|
|
"reference_ready": true
|
|
}
|
|
}
|
|
},
|
|
"calibration": {
|
|
"status": "calibrated",
|
|
"tiers": {
|
|
"low": {
|
|
"algorithm": "predictive-source-path-pcg64-v1",
|
|
"result": {
|
|
"status": "target_met",
|
|
"counts": {
|
|
"gpt__screen023": 11,
|
|
"gpt__screen106": 11,
|
|
"gpt__screen114": 10
|
|
},
|
|
"thresholds": {
|
|
"gpt-6-astra": 0.56729343469469,
|
|
"gpt-6-sol": 0.5,
|
|
"gpt-5.6-terra": 0.5,
|
|
"gpt-6-luna": 0.5,
|
|
"other_known_external": 0.5610207135285652
|
|
},
|
|
"correct_target": 0.7,
|
|
"negative_path_coverage": 0.99,
|
|
"threshold_cap": 0.98,
|
|
"seed": 2026092407,
|
|
"per_source_batches": {
|
|
"gpt-6-astra": 10000,
|
|
"gpt-6-sol": 10000,
|
|
"gpt-5.6-terra": 10000,
|
|
"gpt-6-luna": 10000,
|
|
"openai/gpt-5.6-sol": 10000,
|
|
"openai/gpt-5.6-luna": 10000,
|
|
"z-ai/glm-5.3": 10000
|
|
},
|
|
"source_outcomes": {
|
|
"gpt-6-astra": {
|
|
"correct": 10000,
|
|
"wrong": 0,
|
|
"insufficient": 0
|
|
},
|
|
"gpt-6-sol": {
|
|
"correct": 10000,
|
|
"wrong": 0,
|
|
"insufficient": 0
|
|
},
|
|
"gpt-5.6-terra": {
|
|
"correct": 9999,
|
|
"wrong": 0,
|
|
"insufficient": 1
|
|
},
|
|
"gpt-6-luna": {
|
|
"correct": 9999,
|
|
"wrong": 0,
|
|
"insufficient": 1
|
|
},
|
|
"openai/gpt-5.6-sol": {
|
|
"correct": 9991,
|
|
"wrong": 0,
|
|
"insufficient": 9
|
|
},
|
|
"openai/gpt-5.6-luna": {
|
|
"correct": 7226,
|
|
"wrong": 290,
|
|
"insufficient": 2484
|
|
},
|
|
"z-ai/glm-5.3": {
|
|
"correct": 10000,
|
|
"wrong": 0,
|
|
"insufficient": 0
|
|
}
|
|
},
|
|
"partial_verification": {
|
|
"minimum_sample_ratio": 0.6,
|
|
"same_answer_prefixes": true,
|
|
"patterns": [
|
|
"round_robin",
|
|
"missing20",
|
|
"delayed:gpt__screen023",
|
|
"delayed:gpt__screen106",
|
|
"delayed:gpt__screen114",
|
|
"shuffled"
|
|
],
|
|
"batches_per_source": 10000,
|
|
"max_path_wrong": 0.12,
|
|
"cases": [
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 8687,
|
|
"wrong": 0,
|
|
"strong": 8687
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 6,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 8774,
|
|
"wrong": 3,
|
|
"strong": 8774
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 1,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 1,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 8748,
|
|
"wrong": 1,
|
|
"strong": 8746
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 1,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 8742,
|
|
"wrong": 1,
|
|
"strong": 8742
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 9999
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 9999
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 2,
|
|
"strong": 9999
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 8749,
|
|
"wrong": 3,
|
|
"strong": 8736
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 9994
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 9994
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 9997
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 9995
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 1098,
|
|
"strong": 9489
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 8717,
|
|
"wrong": 757,
|
|
"strong": 7575
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 683,
|
|
"strong": 8979
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 361,
|
|
"strong": 8106
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 510,
|
|
"strong": 7736
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 966,
|
|
"strong": 9458
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 8730,
|
|
"wrong": 0,
|
|
"strong": 8730
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
}
|
|
]
|
|
},
|
|
"scope": "Existing empirical held resampling after user-requested GPT-5.5 reference removal. Not independent validation or live endpoint accuracy. GPT-5.5 is no longer an identification target. Legacy all-source envelope retained; directional gates separately verified."
|
|
},
|
|
"contract": "678a471e11c96670723935d459bc320829399130b4cbcb18c4cf187dc2b6f7fd"
|
|
},
|
|
"medium": {
|
|
"algorithm": "predictive-source-path-pcg64-v1",
|
|
"result": {
|
|
"status": "target_met",
|
|
"counts": {
|
|
"gpt__screen023": 22,
|
|
"gpt__screen106": 21,
|
|
"gpt__screen114": 21
|
|
},
|
|
"thresholds": {
|
|
"gpt-6-astra": 0.5161595114439284,
|
|
"gpt-6-sol": 0.5,
|
|
"gpt-5.6-terra": 0.5,
|
|
"gpt-6-luna": 0.5,
|
|
"other_known_external": 0.51
|
|
},
|
|
"correct_target": 0.99,
|
|
"negative_path_coverage": 0.99,
|
|
"threshold_cap": 0.98,
|
|
"seed": 2026092407,
|
|
"per_source_batches": {
|
|
"gpt-6-astra": 10000,
|
|
"gpt-6-sol": 10000,
|
|
"gpt-5.6-terra": 10000,
|
|
"gpt-6-luna": 10000,
|
|
"openai/gpt-5.6-sol": 10000,
|
|
"openai/gpt-5.6-luna": 10000,
|
|
"z-ai/glm-5.3": 10000
|
|
},
|
|
"source_outcomes": {
|
|
"gpt-6-astra": {
|
|
"correct": 10000,
|
|
"wrong": 0,
|
|
"insufficient": 0
|
|
},
|
|
"gpt-6-sol": {
|
|
"correct": 10000,
|
|
"wrong": 0,
|
|
"insufficient": 0
|
|
},
|
|
"gpt-5.6-terra": {
|
|
"correct": 10000,
|
|
"wrong": 0,
|
|
"insufficient": 0
|
|
},
|
|
"gpt-6-luna": {
|
|
"correct": 10000,
|
|
"wrong": 0,
|
|
"insufficient": 0
|
|
},
|
|
"openai/gpt-5.6-sol": {
|
|
"correct": 10000,
|
|
"wrong": 0,
|
|
"insufficient": 0
|
|
},
|
|
"openai/gpt-5.6-luna": {
|
|
"correct": 9948,
|
|
"wrong": 13,
|
|
"insufficient": 39
|
|
},
|
|
"z-ai/glm-5.3": {
|
|
"correct": 10000,
|
|
"wrong": 0,
|
|
"insufficient": 0
|
|
}
|
|
},
|
|
"partial_verification": {
|
|
"minimum_sample_ratio": 0.6,
|
|
"same_answer_prefixes": true,
|
|
"patterns": [
|
|
"round_robin",
|
|
"missing20",
|
|
"delayed:gpt__screen023",
|
|
"delayed:gpt__screen106",
|
|
"delayed:gpt__screen114",
|
|
"shuffled"
|
|
],
|
|
"batches_per_source": 10000,
|
|
"max_path_wrong": 0.03,
|
|
"cases": [
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 9533,
|
|
"wrong": 0,
|
|
"strong": 9533
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 9480,
|
|
"wrong": 0,
|
|
"strong": 9480
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 9502,
|
|
"wrong": 0,
|
|
"strong": 9502
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 9521,
|
|
"wrong": 0,
|
|
"strong": 9521
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 9561,
|
|
"wrong": 0,
|
|
"strong": 9561
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 169,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 9519,
|
|
"wrong": 106,
|
|
"strong": 9508
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 69,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 18,
|
|
"strong": 9981
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 50,
|
|
"strong": 9996
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 158,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 9533,
|
|
"wrong": 0,
|
|
"strong": 9533
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
}
|
|
]
|
|
},
|
|
"scope": "Existing empirical held resampling after user-requested GPT-5.5 reference removal. Not independent validation or live endpoint accuracy. GPT-5.5 is no longer an identification target. Legacy all-source envelope retained; directional gates separately verified."
|
|
},
|
|
"contract": "cf40c5279da798e014de8398f55904d032c50a12e1c660396d1cde68a9de16b5"
|
|
},
|
|
"high": {
|
|
"algorithm": "predictive-source-path-pcg64-v1",
|
|
"result": {
|
|
"status": "target_met",
|
|
"counts": {
|
|
"gpt__screen023": 43,
|
|
"gpt__screen106": 43,
|
|
"gpt__screen114": 42
|
|
},
|
|
"thresholds": {
|
|
"gpt-6-astra": 0.5,
|
|
"gpt-6-sol": 0.5,
|
|
"gpt-5.6-terra": 0.5,
|
|
"gpt-6-luna": 0.5,
|
|
"other_known_external": 0.5
|
|
},
|
|
"correct_target": 0.99,
|
|
"negative_path_coverage": 0.99,
|
|
"threshold_cap": 0.98,
|
|
"seed": 2026092407,
|
|
"per_source_batches": {
|
|
"gpt-6-astra": 10000,
|
|
"gpt-6-sol": 10000,
|
|
"gpt-5.6-terra": 10000,
|
|
"gpt-6-luna": 10000,
|
|
"openai/gpt-5.6-sol": 10000,
|
|
"openai/gpt-5.6-luna": 10000,
|
|
"z-ai/glm-5.3": 10000
|
|
},
|
|
"source_outcomes": {
|
|
"gpt-6-astra": {
|
|
"correct": 10000,
|
|
"wrong": 0,
|
|
"insufficient": 0
|
|
},
|
|
"gpt-6-sol": {
|
|
"correct": 10000,
|
|
"wrong": 0,
|
|
"insufficient": 0
|
|
},
|
|
"gpt-5.6-terra": {
|
|
"correct": 10000,
|
|
"wrong": 0,
|
|
"insufficient": 0
|
|
},
|
|
"gpt-6-luna": {
|
|
"correct": 10000,
|
|
"wrong": 0,
|
|
"insufficient": 0
|
|
},
|
|
"openai/gpt-5.6-sol": {
|
|
"correct": 10000,
|
|
"wrong": 0,
|
|
"insufficient": 0
|
|
},
|
|
"openai/gpt-5.6-luna": {
|
|
"correct": 10000,
|
|
"wrong": 0,
|
|
"insufficient": 0
|
|
},
|
|
"z-ai/glm-5.3": {
|
|
"correct": 10000,
|
|
"wrong": 0,
|
|
"insufficient": 0
|
|
}
|
|
},
|
|
"partial_verification": {
|
|
"minimum_sample_ratio": 0.6,
|
|
"same_answer_prefixes": true,
|
|
"patterns": [
|
|
"round_robin",
|
|
"missing20",
|
|
"delayed:gpt__screen023",
|
|
"delayed:gpt__screen106",
|
|
"delayed:gpt__screen114",
|
|
"shuffled"
|
|
],
|
|
"batches_per_source": 10000,
|
|
"max_path_wrong": 0.03,
|
|
"cases": [
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 9970,
|
|
"wrong": 0,
|
|
"strong": 9970
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-astra",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 9966,
|
|
"wrong": 0,
|
|
"strong": 9966
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-sol",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 9965,
|
|
"wrong": 0,
|
|
"strong": 9965
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-5.6-terra",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 9961,
|
|
"wrong": 0,
|
|
"strong": 9961
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "gpt-6-luna",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 9965,
|
|
"wrong": 0,
|
|
"strong": 9965
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-sol",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 1,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 9964,
|
|
"wrong": 2,
|
|
"strong": 9964
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 1,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "openai/gpt-5.6-luna",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 4,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "round_robin",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "missing20",
|
|
"batches": 10000,
|
|
"qualified": 9963,
|
|
"wrong": 0,
|
|
"strong": 9963
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "delayed:gpt__screen023",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "delayed:gpt__screen106",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "delayed:gpt__screen114",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
},
|
|
{
|
|
"scope": "check",
|
|
"source": "z-ai/glm-5.3",
|
|
"pattern": "shuffled",
|
|
"batches": 10000,
|
|
"qualified": 10000,
|
|
"wrong": 0,
|
|
"strong": 10000
|
|
}
|
|
]
|
|
},
|
|
"scope": "Existing empirical held resampling after user-requested GPT-5.5 reference removal. Not independent validation or live endpoint accuracy. GPT-5.5 is no longer an identification target. Legacy all-source envelope retained; directional gates separately verified."
|
|
},
|
|
"contract": "d76e468538ed3a070ff893b969a4400b482889f617de4d40185666fa9c90d743"
|
|
}
|
|
}
|
|
},
|
|
"validation": {
|
|
"status": "local_acceptance_candidate",
|
|
"independent_blind": false,
|
|
"scope": "Existing empirical held resampling after user-requested GPT-5.5 reference removal. Not independent validation or live endpoint accuracy. GPT-5.5 is no longer an identification target.",
|
|
"excluded_business_pair": [
|
|
"openai/gpt-5.6-luna",
|
|
"openai/gpt-6-luna"
|
|
],
|
|
"current_full_correct_target": {
|
|
"low": 0.97,
|
|
"medium": 0.99,
|
|
"high": 0.99
|
|
},
|
|
"directional_band_limits": {
|
|
"all": 0.03,
|
|
"early": 0.03,
|
|
"late": 0.01,
|
|
"full": 0.01
|
|
},
|
|
"observed_directional": {
|
|
"low": {
|
|
"current_full_min": 0.9999,
|
|
"rates": {
|
|
"all": 0.0006,
|
|
"early": 0.0006,
|
|
"late": 0.0001,
|
|
"full": 0.0
|
|
}
|
|
},
|
|
"medium": {
|
|
"current_full_min": 1.0,
|
|
"rates": {
|
|
"all": 0.0,
|
|
"early": 0.0,
|
|
"late": 0.0,
|
|
"full": 0.0
|
|
}
|
|
},
|
|
"high": {
|
|
"current_full_min": 1.0,
|
|
"rates": {
|
|
"all": 0.0,
|
|
"early": 0.0,
|
|
"late": 0.0,
|
|
"full": 0.0
|
|
}
|
|
}
|
|
},
|
|
"legacy_contract_note": "Raw symmetric counts are preserved. GPT quick-tier legacy envelope is 70% correct and 12% any-path error across ALL sources (including old Luna); it is NOT the current-model target. Separate directional gates above retain 97% current-model quick target and 99% medium/high targets. The old-Luna/new-Luna exception does not relabel errors as correct.",
|
|
"evidence": [
|
|
{
|
|
"file": "FROZEN.json",
|
|
"sha256": "0baa65a9afcc84880341d849942bed46214d35afd3f55a5a34a348a3f8435894"
|
|
},
|
|
{
|
|
"file": "BUSINESS_AUDIT.json",
|
|
"sha256": "6f045fc8918ab21c6ac66649ae1fea2511a12076d633f493f9e84f0d89e37cb7"
|
|
},
|
|
{
|
|
"file": "BUSINESS_REFINEMENT_CHECK.json",
|
|
"sha256": "2f5af8cd4be03c9b7195f9cf8939ae135b470908e1aaa2af4638f6090fe6571e"
|
|
},
|
|
{
|
|
"file": "SELECTED.json",
|
|
"sha256": "2dbec0ca45e7b08592577b39c917322f311f743040ba70ebf7e03f7dc6db3988"
|
|
},
|
|
{
|
|
"file": "checks.json",
|
|
"sha256": "b3713441971d7a90a5668dcd049fd2b40d3fc04958a286361d80054b101444f9"
|
|
}
|
|
],
|
|
"removed_reference": "openai/gpt-5.5"
|
|
},
|
|
"content_sha256": "e5dee4ccbc9cfd03883c49b3571815f44a4bf1b792730f7918d62c5eba4711e5"
|
|
}
|