{
  "method": "Independent serial, unpadded full-logit forwards; same raw tokens/prefix/context; compare each document to stored batched BPB. Absolute per-document BPB tolerance 1e-5.",
  "models": [
    {
      "model": "HuggingFaceTB/SmolLM2-135M-Instruct",
      "revision": "12fd25f77366fa6b3b4b768ec3050bf629380bac",
      "run_id": "2cd85616f1b932833b5d2618",
      "runner_sha256": "458e80b51765480304e062a2ee97dbeea6177055e9a4a70ba9711cc05d0f288a",
      "documents": 20,
      "max_absolute_document_bpb_difference": 4.367335414309438e-07,
      "passed": true
    },
    {
      "model": "HuggingFaceTB/SmolLM2-360M-Instruct",
      "revision": "a10cc1512eabd3dde888204e902eca88bddb4951",
      "run_id": "5f61f6ffce72b907ddfb6834",
      "runner_sha256": "458e80b51765480304e062a2ee97dbeea6177055e9a4a70ba9711cc05d0f288a",
      "documents": 20,
      "max_absolute_document_bpb_difference": 2.471398746841144e-07,
      "passed": true
    },
    {
      "model": "Qwen/Qwen2.5-0.5B-Instruct",
      "revision": "7ae557604adf67be50417f59c2c2f167def9a775",
      "run_id": "8c739040e87ae8d5e5c8aea2",
      "runner_sha256": "458e80b51765480304e062a2ee97dbeea6177055e9a4a70ba9711cc05d0f288a",
      "documents": 20,
      "max_absolute_document_bpb_difference": 2.5447976561870747e-07,
      "passed": true
    },
    {
      "model": "SlayerLab/Slayer149-ARC",
      "revision": "0e6793ecfb8501dccc1c97dcd944731748eb8d17",
      "run_id": "25fa7b03bb2b9c8422ec385a",
      "runner_sha256": "458e80b51765480304e062a2ee97dbeea6177055e9a4a70ba9711cc05d0f288a",
      "documents": 20,
      "max_absolute_document_bpb_difference": 0.0,
      "passed": true
    },
    {
      "model": "SlayerLab/Slayer149",
      "revision": "7857dd28ec052cd943edcdc1f9e35f7b21dc4f3e",
      "run_id": "6f6b63a1138d8e2ea4f14aa3",
      "runner_sha256": "458e80b51765480304e062a2ee97dbeea6177055e9a4a70ba9711cc05d0f288a",
      "documents": 20,
      "max_absolute_document_bpb_difference": 0.0,
      "passed": true
    },
    {
      "model": "SlayerLab/Slayer149-Balanced-ARC-E-ft-45500",
      "revision": "e7d4b8dea44125264077c2ce6d2f141379f999cd",
      "run_id": "91f182a4e98dba674fa0f1e0",
      "runner_sha256": "458e80b51765480304e062a2ee97dbeea6177055e9a4a70ba9711cc05d0f288a",
      "documents": 20,
      "max_absolute_document_bpb_difference": 0.0,
      "passed": true
    },
    {
      "model": "SlayerLab/Slayer139-1.01",
      "revision": "2e72c696d7c5cad73dc58606b0751b911dc94495",
      "run_id": "4226c308d87b13fbba56f017",
      "runner_sha256": "458e80b51765480304e062a2ee97dbeea6177055e9a4a70ba9711cc05d0f288a",
      "documents": 20,
      "max_absolute_document_bpb_difference": 0.0,
      "passed": true
    },
    {
      "model": "SlayerLab/Slayer139-1.01b",
      "revision": "283ffb6d620be2d6ea1f2429d721962dc8e73633",
      "run_id": "24deee06e103885ddbf5a347",
      "runner_sha256": "458e80b51765480304e062a2ee97dbeea6177055e9a4a70ba9711cc05d0f288a",
      "documents": 20,
      "max_absolute_document_bpb_difference": 0.0,
      "passed": true
    },
    {
      "model": "SlayerLab/Slayer149-Balanced-ARC-E-ft-v2",
      "revision": "7d11e8fd63764baf00c276f586b2aa776648ed33",
      "run_id": "467e61501a76665648a52a2f",
      "runner_sha256": "458e80b51765480304e062a2ee97dbeea6177055e9a4a70ba9711cc05d0f288a",
      "documents": 20,
      "max_absolute_document_bpb_difference": 0.0,
      "passed": true
    },
    {
      "model": "SlayerLab/Slayer139-1.02",
      "revision": "c4298d757ba39cba19dd768777da9fadf319ff61",
      "run_id": "16708cfd8324d1464b5c7e78",
      "runner_sha256": "458e80b51765480304e062a2ee97dbeea6177055e9a4a70ba9711cc05d0f288a",
      "documents": 20,
      "max_absolute_document_bpb_difference": 0.0,
      "passed": true
    },
    {
      "model": "SlayerLab/Slayer-149M-Decision-v1",
      "revision": "ccfd7d3d91f3894beb1cd096b5d067333ba3d4c7",
      "run_id": "24caa458a5e2ad031f93eaee",
      "runner_sha256": "458e80b51765480304e062a2ee97dbeea6177055e9a4a70ba9711cc05d0f288a",
      "documents": 20,
      "max_absolute_document_bpb_difference": 0.0,
      "passed": true
    }
  ],
  "finished": 1792198733,
  "reference_source_sha256": "af2653d0cda1a3db7420a8d0684cb83a7076c301e91c50458ff269424cf27885",
  "scope_note": "Independent check of the initial 2026-10 cohort. Run IDs identify the validated snapshot; future monthly refreshes are separate evidence."
}
