76 lines
3.0 KiB
JSON
76 lines
3.0 KiB
JSON
{
|
|
"schemaVersion": 1,
|
|
"reviewId": "aas-v1-baseline-reviewer-beta",
|
|
"reviewedAt": "2026-07-17",
|
|
"reviewer": {
|
|
"identity": "codex-independent-beta",
|
|
"role": "independent-baseline-reviewer",
|
|
"implementedScorer": false,
|
|
"authoredReviewedArtifacts": false,
|
|
"reliedOnReviewerAlpha": false
|
|
},
|
|
"overall": {
|
|
"decision": "approved",
|
|
"reasonCodes": [
|
|
"AAS_REVIEW_HELD_OUT_APPROVED",
|
|
"AAS_REVIEW_TUNING_APPROVED",
|
|
"AAS_REVIEW_ABSTENTION_APPROVED"
|
|
],
|
|
"scope": "baseline-corpus-only"
|
|
},
|
|
"methodReasonCodes": [
|
|
"AAS_REVIEW_GRID_AND_PAIRING_VERIFIED",
|
|
"AAS_REVIEW_GOALS_AND_REQUIRES_SKILL_VERIFIED",
|
|
"AAS_REVIEW_GOLD_EQUIVALENTS_VERIFIED",
|
|
"AAS_REVIEW_CATALOG_AND_POLICY_VERIFIED",
|
|
"AAS_REVIEW_NO_ARTIFICIAL_LABEL_LEAK",
|
|
"AAS_REVIEW_NO_RAW_CODE_SECRET_OR_ABSOLUTE_PATH",
|
|
"AAS_REVIEW_TASK_FAMILIES_UNIQUE",
|
|
"AAS_REVIEW_TUNING_HELDOUT_SEPARATED",
|
|
"AAS_REVIEW_ABSTENTION_OOD_VERIFIED"
|
|
],
|
|
"warnings": [
|
|
{
|
|
"code": "AAS_REVIEW_METADATA_EVIDENCE_CONTRACT_PENDING",
|
|
"details": {
|
|
"uniqueHeldOutGoldSkillIds": 170,
|
|
"legacyCatalogHasV1JudgmentEvidenceRefs": false
|
|
},
|
|
"constraint": "Baseline approval does not make legacy metadata evidence-backed. Unknown remains unknown; candidates stay in discovery until v1 structured evidence exists."
|
|
}
|
|
],
|
|
"digestContract": {
|
|
"pairPayload": {
|
|
"heldOut": "recursively key-sorted JSON.stringify({case, gold}) with gold.reviews replaced by []",
|
|
"tuning": "recursively key-sorted JSON.stringify({case, gold}) with gold.reviews replaced by []",
|
|
"abstention": "recursively key-sorted JSON.stringify({case, label}) with label.reviews replaced by []"
|
|
},
|
|
"pairDigest": "sha256- prefix followed by lowercase SHA-256 hex of the UTF-8 pair payload",
|
|
"pairManifest": "JSON.stringify of all [caseId,pairDigest] tuples sorted by caseId ascending using JavaScript code-unit order",
|
|
"pairManifestRootDigest": "sha256- prefix followed by lowercase SHA-256 hex of the UTF-8 pair manifest",
|
|
"globalManifest": "the same tuple form and ordering across all three splits"
|
|
},
|
|
"pairCount": 270,
|
|
"pairManifestRootDigest": "sha256-91a89667c3d178b2a2dfe65236cdbea70a2152bf93efd74cfb05f06041620939",
|
|
"splits": {
|
|
"heldOut": {
|
|
"decision": "approved",
|
|
"reasonCode": "AAS_REVIEW_PAIR_APPROVED",
|
|
"pairCount": 180,
|
|
"pairManifestRootDigest": "sha256-b68254bf102ff041b6b32dac2aad244e726ca42118d17885c146965b299a3e16"
|
|
},
|
|
"tuning": {
|
|
"decision": "approved",
|
|
"reasonCode": "AAS_REVIEW_TUNING_PAIR_APPROVED",
|
|
"pairCount": 60,
|
|
"pairManifestRootDigest": "sha256-6a77ed03d82ba9dc22b4e7df01147f174161543c4cb863cefb77fbdd92544ff8"
|
|
},
|
|
"abstention": {
|
|
"decision": "approved",
|
|
"reasonCode": "AAS_REVIEW_ABSTENTION_PAIR_APPROVED",
|
|
"pairCount": 30,
|
|
"pairManifestRootDigest": "sha256-b38198df8e156459f3a43d55adb0057513cc9dbe71c078a2de6484810fc2c060"
|
|
}
|
|
}
|
|
}
|