43 lines
3.4 KiB
JSON
43 lines
3.4 KiB
JSON
{
|
|
"actor": "research-abi-independent",
|
|
"artifacts": [
|
|
{
|
|
"path": "verify/results/research-completion-abi-independent/verification-plan-run-d94d4516b9d898793318805e.md",
|
|
"sha256": "78b7dea4e916d5ef7e4b10b488970e51407094580bbc4d2f807082fb0018a76d"
|
|
},
|
|
{
|
|
"path": "verify/results/research-completion-abi-independent/reproduce_run_d94d4516b9d898793318805e.py",
|
|
"sha256": "cd71d370d9e1c393a98f6c1034f8e785b0b3e95a2ecb96319ee193b23f656eaa"
|
|
},
|
|
{
|
|
"path": "verify/results/research-completion-abi-independent/heldout_run_d94d4516b9d898793318805e.py",
|
|
"sha256": "14935c800bac1dc60dce71fa8067b9c0a282d2871002381a4432b0141be8beec"
|
|
},
|
|
{
|
|
"path": "verify/results/research-completion-abi-independent/run-d94d4516b9d898793318805e/manifest.json",
|
|
"sha256": "c7313047b4f013068649fc7784db4db277a6ca42083ec5cdb446221559561deb"
|
|
},
|
|
{
|
|
"path": "verify/results/research-completion-abi-independent/run-d94d4516b9d898793318805e/independent-state.json",
|
|
"sha256": "6463816f322803b2b7b479e1fc116c4ffa33fb2772a3005d492422a00acea433"
|
|
},
|
|
{
|
|
"path": "verify/results/research-completion-abi-independent/run-d94d4516b9d898793318805e/heldout-boundary.json",
|
|
"sha256": "c7c950672d987d2243bfeede543f6ff37918c806873def6e9c3bd12ba5c5f091"
|
|
},
|
|
{
|
|
"path": "verify/results/research-completion-abi-independent/result-run-d94d4516b9d898793318805e.md",
|
|
"sha256": "b2278ae463654bf05f8456fad5c14dd837539cc1d5f0840f5a2098026a17e321"
|
|
}
|
|
],
|
|
"basis": "aaddeb76f74bbd1a43671f882770ec84a7d690b4060d060ef1a68f918e8edbb0",
|
|
"contract": "research-completion-abi",
|
|
"id": "caa4abb2f50ba575121c1b6a",
|
|
"model": "openai/gpt-5.6-sol",
|
|
"next_action": "Record a fresh independent PASS verdict over the current integrated evidence array, validate it, then checkpoint quantum end.",
|
|
"role": "verifier",
|
|
"schema": "sots-checkpoint/1",
|
|
"session": "run-d94d4516b9d898793318805e",
|
|
"summary": "POST-EXPERIMENT PRE-VERDICT CHECKPOINT. OBSERVATIONS: fresh handoff reproduction exited 0 and produced 22 nonempty objdump windows plus strict save-reader/checksum (24 positive subprocesses, empty stderr, no skips). Fresh manifest c7313047b4f013068649fc7784db4db277a6ca42083ec5cdb446221559561deb equals integrated manifest except session; state artifact 6463816f322803b2b7b479e1fc116c4ffa33fb2772a3005d492422a00acea433 is byte-identical. Held-out direct-PE minus-one/plus-one boundary ablation passed all seven exact and fourteen shifted negative controls; artifact c7c950672d987d2243bfeede543f6ff37918c806873def6e9c3bd12ba5c5f091. Required 0x2c/0x74 stride, deep-copy, three-delete, description, three unordered-or-unequal float branches exposed. Actual state/RNG independently checked: EvNxID 4, turn-3 count 2, complete event ID 3, exact 609080-byte rebuild, RNG leaf 2503 bytes/digest 0978fdf34ff7962f76c2de810dc93e0a. 37 campaign tests pass; campaign validate passes. Canonical source binding still matches integrated evidence: engine ccd8e02083e8d2e2b3e97976ace2273c8f924dfc02a39e919004eaf3544c50fd, RE e74ef6cff202ee35a6fdc38d63842b4826a074e34502dfc2981a33ffea912bd4. DECISION: scoped integrated PASS is supported for static reproduction plus one archived save only. No new surprise/prediction failure. Historical provenance/description failures remain recorded. BLOCKERS/RESIDUALS: no live capacity/string/duplicate/NaN/unwind/allocator-family/original differential/replay exposure.",
|
|
"timestamp": "2026-09-10T01:39:33.390227+00:00"
|
|
}
|