{
 "backend_profile": {
  "attention": "eager",
  "device": "cuda",
  "dtype": "bfloat16",
  "recurrent_state_dtype": "float32"
 },
 "batch": {
  "bytes": 381381,
  "note": "The exact file copied from the miner object at the turn (before the batch-cap trim).",
  "published": true,
  "sha256": "e1def164a7048cf25bf89d9fb0459e45c49cec1ae546b8d2125c1239e22098f8",
  "url": "/k1/turns/354/batch.json"
 },
 "candidate": {
  "files_sha256": {
   "chat_template.jinja": "59a64ebb4df6d1489d09a91267cf3ceb106162d4a893c4f84833cfb8c897ff63",
   "config.json": "44a011dc9cc7203c8b13867f643ef12b7056b2eb8bbf9410704e30a2cec27928",
   "generation_config.json": "cde05246e60160302c49ac904fbc45365d4a98b16e003b4a1ee15cfc7315bd92",
   "model-00001-of-00010.safetensors": "568dfa79c9721b48f96fa04173032fc2c49872e9c61438591cbf40aa284f608e",
   "model-00002-of-00010.safetensors": "1d376e16ecdbb3fa042d91dc0c2ae675bef392e33d0a146e5099fdb7f717f2e4",
   "model-00003-of-00010.safetensors": "d3592ec64600bd43d56f36ee8b221621e6fafd9d59e4994d298c367a424818f3",
   "model-00004-of-00010.safetensors": "5570b73ec6e302d8fcda2395328f0ea12a9e975cc23f80b7e35ad4c4b53285e9",
   "model-00005-of-00010.safetensors": "48dd2c4e399ab275c39a3453551b604ba475f6c926ca01dfdc125103f34a0c11",
   "model-00006-of-00010.safetensors": "7a3a9cb1360092e8e8ceee19e66ddf6b152df094278c68fc66d051cb4e22b1e2",
   "model-00007-of-00010.safetensors": "e3f3cae7c17759a21b0018e92dc9390709e9ed4fe5d9a00f17dfb73404394692",
   "model-00008-of-00010.safetensors": "f596227272fe538f306a91e675d57c986b4154f1ee4287f61082e7ebf4415c54",
   "model-00009-of-00010.safetensors": "26b28b8e7f18c492017dd189bc45bddf6a5bd0348b0ed4bb6f9103e6514e57bf",
   "model-00010-of-00010.safetensors": "e981496fcd8c840e169e5d36a6a73e5839a2edfb48fdd324f8898d6f350ab135",
   "model.safetensors.index.json": "e6b1277a7962e812d1a3f6773ac6c55cc32f203b9af71e2d320e482b744d7608",
   "tokenizer.json": "06b9509352d2af50381ab2247e083b80d32d5c0aba91c272ca9ff729b6a0e523",
   "tokenizer_config.json": "7488550a894dcf741616126e322ea31d66cf9457b59ec5825f4c7efdced421f8"
  },
  "id": "ce3971dc21a17a964c03273b68a969984dd43503094937e56bee8cf1280b1853",
  "note": "GREEN candidates become the next head; RED candidates are discarded but their hashes are listed here.",
  "promoted": true,
  "r2_prefix": "mimo/affine-mimo9b-grpo-20261013/checkpoints/ce3971dc21a17a964c03273b68a969984dd43503094937e56bee8cf1280b1853/model/"
 },
 "decision": {
  "batch_tasks": 0,
  "candidate": "ce3971dc21a17a964c03273b68a969984dd43503094937e56bee8cf1280b1853",
  "cap_max": 0.2,
  "commitment": "a36b4dfd92d41d82089f4b4035a7e9beac5562b0de5e2286420c085d8e57b840",
  "decided_at": 1793453966.4619577,
  "decided_at_cr": "2026-10-31 07:39:26 CR (UTC-6)",
  "decision": "green",
  "evaluated": {
   "c_cap_rate": 0.10049999999999999,
   "delta": 0.08360008536064877,
   "redacted": "held-out round open: n, pass rates on the paired tasks, gains/losses and batch overlap are published with the round reveal (they would leak which tasks are in the secret set)",
   "z": 3.8804607275736713
  },
  "head": "87fa740811e75e4c221258ce6cf2d8cb1db1519d69aa0a65b42cb5fddea5a499",
  "min_tasks": 100,
  "mode": "live",
  "reasons": [],
  "sample_size": 179,
  "z_inclusive": false,
  "z_min": 1.0
 },
 "dispatch": {
  "envelope_file_sha256": "fcfed358bae5cca41e84217713beaf2a7d2a1d1898d2ed28c488ee022dc3d700",
  "header": {
   "checkpoint": "87fa740811e75e4c221258ce6cf2d8cb1db1519d69aa0a65b42cb5fddea5a499",
   "epoch": "affine-mimo9b-grpo-20261013-e000354",
   "format": "affine-batch-v1",
   "manifest_sha256": "a9c6cc0406d0b0186bbdc0283a6ab4759191ac5046e94bac177d984d61239855",
   "run_id": "affine-mimo9b-grpo-20261013",
   "selected_rollout_count": 125,
   "selected_task_count": 125,
   "version": "sn120-ff-dispatch-v1"
  },
  "selected": [
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/f51727fb52fddbaecaafcc5112bb0812ce1dd550d3fd645ede58972af956be3f.json",
     "sha256": "f51727fb52fddbaecaafcc5112bb0812ce1dd550d3fd645ede58972af956be3f",
     "size": 2303
    },
    "index": 0,
    "prompt_sha256": "042a3ff560ad5dc0d6207f2bdd1bf27e7c22507ca63571623d6c8a3502f8ed8c",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/cdae5c9b95d55580d21be24f7cec74cb11a6a93f6748723cfedde273a7d0f51f.json",
     "sha256": "cdae5c9b95d55580d21be24f7cec74cb11a6a93f6748723cfedde273a7d0f51f",
     "size": 4080
    },
    "index": 1,
    "prompt_sha256": "15b867aa8c69c153e502c62fedfed21471c954b6363adce748b5d3b0c2fe2c41",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/a2837fe3ca4f53360c7a1620d79c32fdffd043a311d26ff27ff020af75335be6.json",
     "sha256": "a2837fe3ca4f53360c7a1620d79c32fdffd043a311d26ff27ff020af75335be6",
     "size": 6107
    },
    "index": 2,
    "prompt_sha256": "d2a1c5d8d0cf829ad8655ebf03ae27602df42c1321a037616f28381d368f2c55",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/62f0164603dbfa981995fd345b78383168713b56116e4c912750bc45ebe14fd7.json",
     "sha256": "62f0164603dbfa981995fd345b78383168713b56116e4c912750bc45ebe14fd7",
     "size": 3141
    },
    "index": 3,
    "prompt_sha256": "802597d7d74dbb2db54d0e60b1a309f0acda1eee59cb6a4861b6fb1b3b4362f9",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/76574e4c6026d87c9ec8b5b0753af9ba98efc69e9b2c15f44b56718ac5b16a2b.json",
     "sha256": "76574e4c6026d87c9ec8b5b0753af9ba98efc69e9b2c15f44b56718ac5b16a2b",
     "size": 7398
    },
    "index": 4,
    "prompt_sha256": "2e4b12e815447a89dc296817fe17142012c3bef1b41eb0ea3000174184a8a9e8",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/7ed38c9c325200599e732327d4c258f475c2d7994dc36a61ba6e981f9e3cc3b4.json",
     "sha256": "7ed38c9c325200599e732327d4c258f475c2d7994dc36a61ba6e981f9e3cc3b4",
     "size": 3416
    },
    "index": 5,
    "prompt_sha256": "2f818a88717eaa623777c082b01cafa130623c2cc8f6b6ff2e8feb32e8026882",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/d7d17185ca4e390aa57f16bb20a7648dbdd516ef542e747fa0f38157e314daf9.json",
     "sha256": "d7d17185ca4e390aa57f16bb20a7648dbdd516ef542e747fa0f38157e314daf9",
     "size": 3203
    },
    "index": 6,
    "prompt_sha256": "dde212ee1844b8918d751fc23f19356347b9c9f124322572cfcd806df7cb5711",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/51fa4058c6af3b4963e5bebcb7d2ec2ac6a87e762fc897684547f778e0dea68c.json",
     "sha256": "51fa4058c6af3b4963e5bebcb7d2ec2ac6a87e762fc897684547f778e0dea68c",
     "size": 2135
    },
    "index": 7,
    "prompt_sha256": "124e4dd24b8dee8e97d64ca6796fae93cf087b6f4ff020c31a667cf1b8702df4",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/e5010f901dc63185155af840a3fbbbe76fa4886ff323a07985bd402fcbe7a994.json",
     "sha256": "e5010f901dc63185155af840a3fbbbe76fa4886ff323a07985bd402fcbe7a994",
     "size": 3849
    },
    "index": 8,
    "prompt_sha256": "925405c068e7cbc88c9f5ec8b018fb922bbd45b1284217fde42b15ecd0fd7aef",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/831d0fc1352e110befbc1320aa4c13da85ec273ac7273eaac74bbc083684ae34.json",
     "sha256": "831d0fc1352e110befbc1320aa4c13da85ec273ac7273eaac74bbc083684ae34",
     "size": 4039
    },
    "index": 9,
    "prompt_sha256": "2dd7a04ff6e7ef4ddd6d358408ad7d5de92636dc4760b5e88c3cb5682184115c",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/06ec1a059302c317c20e12199ae833ae73a7688dd3e260c835eb5e1b9814dafa.json",
     "sha256": "06ec1a059302c317c20e12199ae833ae73a7688dd3e260c835eb5e1b9814dafa",
     "size": 4908
    },
    "index": 10,
    "prompt_sha256": "3bcc877c1daf5f62feea6e42009477f9f37a0dfacb986a7ca5e0775b2a6c46f9",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/66b5afa49d182228ecec3b774015577a5dc39a3207b8512d3bca877b5ca0b210.json",
     "sha256": "66b5afa49d182228ecec3b774015577a5dc39a3207b8512d3bca877b5ca0b210",
     "size": 13645
    },
    "index": 11,
    "prompt_sha256": "86164a3dc9733c10df19b88dc0a6344655b730f2236c639e7948804eae4ad156",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/cf330ee5ea9fabdb58d30926e710e8360aacc3d0406678f2ba5d937d45f1009a.json",
     "sha256": "cf330ee5ea9fabdb58d30926e710e8360aacc3d0406678f2ba5d937d45f1009a",
     "size": 1915
    },
    "index": 12,
    "prompt_sha256": "bf52adbe059e13f2e7be8750e987380636b1b30a799b7d643c2f58dd2e105894",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/4db533b755971570d6003fd749710d4e8b8cf1a189d3e173ec78a9993d30c8b7.json",
     "sha256": "4db533b755971570d6003fd749710d4e8b8cf1a189d3e173ec78a9993d30c8b7",
     "size": 6210
    },
    "index": 13,
    "prompt_sha256": "1a727d1258871b6ef814f10f4ad7dfa3cea47de8d9d053953ad44a799732044e",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/e18d81d4364db6b1d8b797e4017b1e7116fc6b71eb800077cf20888c5f1170b9.json",
     "sha256": "e18d81d4364db6b1d8b797e4017b1e7116fc6b71eb800077cf20888c5f1170b9",
     "size": 12357
    },
    "index": 14,
    "prompt_sha256": "051daee722f79350e96f070e22f305695ca28f82f425ec41c26244e5eb49f960",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/699d38b1d8ca02fc2f102a06cc72adfcad75343336e206acbd8f909577c06989.json",
     "sha256": "699d38b1d8ca02fc2f102a06cc72adfcad75343336e206acbd8f909577c06989",
     "size": 5323
    },
    "index": 15,
    "prompt_sha256": "71932984962c55b91fba4d97cb6ca26b32079b668ba0835ef0eb9553455e279b",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/43d3b210ab37f9412548d2e636a011ac8bb47ce9718dae61ce4bc61b40e7730d.json",
     "sha256": "43d3b210ab37f9412548d2e636a011ac8bb47ce9718dae61ce4bc61b40e7730d",
     "size": 1715
    },
    "index": 16,
    "prompt_sha256": "be597d397e2e26a1a634127941ddbe836be7a6b1e3b8bd5010b0d81d14b1add0",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/de0c6486933e75c628f1ffb420b2dc3775534a74925bf32fd9d95e5d86a3e6bd.json",
     "sha256": "de0c6486933e75c628f1ffb420b2dc3775534a74925bf32fd9d95e5d86a3e6bd",
     "size": 13315
    },
    "index": 17,
    "prompt_sha256": "ca62f6057e7d113aed3e74583514cb449fef009f30abb810ce4edaa731c87efb",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/7ff6c2b4d2f4eac98bad134ca7df7e9f81b0913333dd585d86358da42b8bd225.json",
     "sha256": "7ff6c2b4d2f4eac98bad134ca7df7e9f81b0913333dd585d86358da42b8bd225",
     "size": 6889
    },
    "index": 18,
    "prompt_sha256": "b12640f61524792fc5ab0f3974bbf25351179268fa2103e48aa39f21b580ab11",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/458263c23337ac96eb3912f51fa81bd6d44dbd4ca4cabf85d8aca89278675087.json",
     "sha256": "458263c23337ac96eb3912f51fa81bd6d44dbd4ca4cabf85d8aca89278675087",
     "size": 5981
    },
    "index": 19,
    "prompt_sha256": "1575a49383e3172bfb299ca6b88105c1f02ff0255cf5995a6f8319f705c385fe",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/c22bb4abfa5e63256990935b9def714486b3e753ee5ea743848dda049c745681.json",
     "sha256": "c22bb4abfa5e63256990935b9def714486b3e753ee5ea743848dda049c745681",
     "size": 2582
    },
    "index": 20,
    "prompt_sha256": "cdcf5e9facddfea8d1c4fca3b8c299b6231879adcd9684618bbc4a1a9a70303e",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/2f89b8f4fe084f6a08efd61132633472202dd4c56ac0b019cb2437ad9bc7b135.json",
     "sha256": "2f89b8f4fe084f6a08efd61132633472202dd4c56ac0b019cb2437ad9bc7b135",
     "size": 1822
    },
    "index": 21,
    "prompt_sha256": "7870d643d092492ce2cead0d6391976e4dfef4c6061ff83aadbd9f30c1af614b",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/ecd627cf58e20d238250c299c72a71a67a49b40e80983a0c4026bf1351caae94.json",
     "sha256": "ecd627cf58e20d238250c299c72a71a67a49b40e80983a0c4026bf1351caae94",
     "size": 3515
    },
    "index": 22,
    "prompt_sha256": "9739e03fc6c1c792889eac5fbcef00c02e32f0388bbdda9d39a119db3900dcad",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/027d24a8af02196273bd291df3660e23eb72fecaa5e897122a3706c1b68bdece.json",
     "sha256": "027d24a8af02196273bd291df3660e23eb72fecaa5e897122a3706c1b68bdece",
     "size": 13785
    },
    "index": 23,
    "prompt_sha256": "feae2703852bbc7691bde9bd223807efe0438d643fa1d641b4c28d31928ef07a",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/7652b3b6580e3900897dd234e60f4e1ac253e4f1cc41b75e30a209a0dace2ea4.json",
     "sha256": "7652b3b6580e3900897dd234e60f4e1ac253e4f1cc41b75e30a209a0dace2ea4",
     "size": 1993
    },
    "index": 24,
    "prompt_sha256": "02c533a0f6a50a0344d2390c6beffd38beb70fabeec153895d227b709a07ed07",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/09527862bd9b6de0b04ce20d18115df2a01c04fea532b05ccb88e729545cbef7.json",
     "sha256": "09527862bd9b6de0b04ce20d18115df2a01c04fea532b05ccb88e729545cbef7",
     "size": 2475
    },
    "index": 25,
    "prompt_sha256": "bdc230cb13fccc8c0493f640244647877af6f978eb73195102568c1b7907b265",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/c9e4ccf0db9641053a3ccc16bad8c89f939d6dcfabf1ba90a7445ae3f8211442.json",
     "sha256": "c9e4ccf0db9641053a3ccc16bad8c89f939d6dcfabf1ba90a7445ae3f8211442",
     "size": 1798
    },
    "index": 26,
    "prompt_sha256": "b40a27089e8fb5f173974e6e500bab4a20d641a9ba50c6816cc39825c87aaf61",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/71465c2fc02d7572756c675e8af1ab3f1fe588f74fbde72daee7ea57cc7eb3d3.json",
     "sha256": "71465c2fc02d7572756c675e8af1ab3f1fe588f74fbde72daee7ea57cc7eb3d3",
     "size": 4954
    },
    "index": 27,
    "prompt_sha256": "79d034f93c2e2a0a9c9dbb2195c18d816175e437088bf441a1870651057dfdbe",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/e2bd9e706778864e19ddb8e29ff6d4d487f6ac55d0d3d1241c42b133fea21dd6.json",
     "sha256": "e2bd9e706778864e19ddb8e29ff6d4d487f6ac55d0d3d1241c42b133fea21dd6",
     "size": 1589
    },
    "index": 28,
    "prompt_sha256": "d4421c689be8404581b435cd6eee1e10f8ae31a279b772882c3ec2d7cdbf9784",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/199036a122b41e4f5ac57593a63272e70b3a81fea4a2ff7b1d87d352c511ce12.json",
     "sha256": "199036a122b41e4f5ac57593a63272e70b3a81fea4a2ff7b1d87d352c511ce12",
     "size": 2110
    },
    "index": 29,
    "prompt_sha256": "2cabde7865509e93e52454ebcc7188d9fba73a45aadb202dad3d68ee7924231d",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/b02b2260dd2f4fdbd79f02beda4874bd27591368d1ab99b88c876ae43badacb3.json",
     "sha256": "b02b2260dd2f4fdbd79f02beda4874bd27591368d1ab99b88c876ae43badacb3",
     "size": 10755
    },
    "index": 30,
    "prompt_sha256": "68f1a241af6db298ba9196e6daca03d92c443b8e9f46e458884cac74efdee635",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/dee6297b0a38f760b6e7e35cc1891eace2f32ac83802bfdff68105b65176069f.json",
     "sha256": "dee6297b0a38f760b6e7e35cc1891eace2f32ac83802bfdff68105b65176069f",
     "size": 2887
    },
    "index": 31,
    "prompt_sha256": "fe71bb99f40d03a8f55f48abfa0d05756e6fff9d3d43084a49a6ce9cc2a16e6e",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/15b768ff9ecfde5eabaa16431599c876ae94e0a14389ba40a06b11d810323261.json",
     "sha256": "15b768ff9ecfde5eabaa16431599c876ae94e0a14389ba40a06b11d810323261",
     "size": 2472
    },
    "index": 32,
    "prompt_sha256": "bd52ced0fcc7b3eca301bc2e50fae965b60c3e44812ceee23f1dc2c46a7fd360",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/d94e1c383417745e9a8e0e851ba906b2fd05945499f8ff6885fcfda89de9f8cc.json",
     "sha256": "d94e1c383417745e9a8e0e851ba906b2fd05945499f8ff6885fcfda89de9f8cc",
     "size": 1734
    },
    "index": 33,
    "prompt_sha256": "72339280161474cd0a7896e14ccb9c4a6746af5f030153610b2ea8ee92f2c516",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/025d1a3fcab927c5adc7a20970593f51c3ef4e0f0feb8aebda16e18bc1614afc.json",
     "sha256": "025d1a3fcab927c5adc7a20970593f51c3ef4e0f0feb8aebda16e18bc1614afc",
     "size": 6839
    },
    "index": 34,
    "prompt_sha256": "90d4d6dca70029d5f3bc5e949e4accf328b99d89531cf05c68bac3e61ed9f411",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/b8216db70db97bf84e244a32983243a9021e5e7c6f3cb3aea94ef999788b238c.json",
     "sha256": "b8216db70db97bf84e244a32983243a9021e5e7c6f3cb3aea94ef999788b238c",
     "size": 2399
    },
    "index": 35,
    "prompt_sha256": "13d460b3166f604b026f0612b4a9a31fed180f4692fad3da422b2325b1a06e32",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/4bf0c80c2e014736129058d85ac6d4b6e22c03041504d47aebfd81db0a76f597.json",
     "sha256": "4bf0c80c2e014736129058d85ac6d4b6e22c03041504d47aebfd81db0a76f597",
     "size": 5096
    },
    "index": 36,
    "prompt_sha256": "3ea1fc888710c35b76e1fd5f3a4de6d7acbd5df038be746c7ddda19092cabf6b",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/255d826bf4a55b26c39ae7e2b261c7e25796fde9f69112c33ad6cc6d123134e9.json",
     "sha256": "255d826bf4a55b26c39ae7e2b261c7e25796fde9f69112c33ad6cc6d123134e9",
     "size": 23352
    },
    "index": 37,
    "prompt_sha256": "b97de8c571099294631bd38e10dc710363e4bb3eded247cb288953d5d82412c2",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/88e6bc2095ad4f359c5bde1a06b5911bf2439e1b5b262c3d8115a34adcade402.json",
     "sha256": "88e6bc2095ad4f359c5bde1a06b5911bf2439e1b5b262c3d8115a34adcade402",
     "size": 3281
    },
    "index": 38,
    "prompt_sha256": "c1e5e4ea4c91f1ecb4338ed48bc95dc69c4611bdf252e999140a3378bdd52669",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/6a023a6665b4427afe118b5ca045bf59b0b884e19cea4af8d8e2d01a8fee1284.json",
     "sha256": "6a023a6665b4427afe118b5ca045bf59b0b884e19cea4af8d8e2d01a8fee1284",
     "size": 1749
    },
    "index": 39,
    "prompt_sha256": "2e57638f73f9cfdd35c5eb9ed08edae962cff9054eeae7f9a068194030331883",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/8735aeee1d1dabe83798a67805c1c1b74c5a6d7f0d8378b996a536a2a6e1595c.json",
     "sha256": "8735aeee1d1dabe83798a67805c1c1b74c5a6d7f0d8378b996a536a2a6e1595c",
     "size": 3397
    },
    "index": 40,
    "prompt_sha256": "e19d6a37b96e555b8a555abb9cfccf29b3196c48e9fdb2bd6e41e25285e42f3e",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/7b8825e53408016a1f01a3abbf625a110e3cedffaafecbeaf4857c7123a47d5e.json",
     "sha256": "7b8825e53408016a1f01a3abbf625a110e3cedffaafecbeaf4857c7123a47d5e",
     "size": 4626
    },
    "index": 41,
    "prompt_sha256": "7bb9a882bda0ec28dddcace9a5de7eff3e428a4a77ce32ccade61d704ca6da59",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/1bbfe6a909ce1700a9ae065d5c28302b89675673e6aa8d23c032fb7a48c3a43f.json",
     "sha256": "1bbfe6a909ce1700a9ae065d5c28302b89675673e6aa8d23c032fb7a48c3a43f",
     "size": 4559
    },
    "index": 42,
    "prompt_sha256": "85ed7d784264f28504c5589c17ec643bf1faf1c0da4a246029163f114a5a60f1",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/906cb7e3326f6b468f34ee3cf4099495c363461d3fb4d939b6700e73641bbcfb.json",
     "sha256": "906cb7e3326f6b468f34ee3cf4099495c363461d3fb4d939b6700e73641bbcfb",
     "size": 3260
    },
    "index": 43,
    "prompt_sha256": "a80674f30b450999ac1dc5c1521c187b86ab7e7444d4df1cd8b3925a073590d4",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/87535617e445332e7bc91c6e457e854a84f034cdb0b2e31ff21a90b6029b9b25.json",
     "sha256": "87535617e445332e7bc91c6e457e854a84f034cdb0b2e31ff21a90b6029b9b25",
     "size": 20830
    },
    "index": 44,
    "prompt_sha256": "3232915dfdae1ac6160315882aa44da80823bc0f2ec132343e6b33de1c253d87",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/5a401a102f9fe80851861a217ca4df71de33fcc7daa4566cbeeff9015960641e.json",
     "sha256": "5a401a102f9fe80851861a217ca4df71de33fcc7daa4566cbeeff9015960641e",
     "size": 9530
    },
    "index": 45,
    "prompt_sha256": "e81c9469390af42df88c7ad45b123d0220e5e5a932e456b507ce2f420417a407",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/9969b33784ea2bb3da21a78a45a3a0d71e7b8ddabb8c00701ccbaab3306ec2fb.json",
     "sha256": "9969b33784ea2bb3da21a78a45a3a0d71e7b8ddabb8c00701ccbaab3306ec2fb",
     "size": 2216
    },
    "index": 46,
    "prompt_sha256": "be62406b7dfdda8b4d6968cff213aec04b12414af25e6be3671d712efc0096f2",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/de94d4b8e7ebe9b636862fbf296201cdd397cd3bde01c58953f6a297546e58b1.json",
     "sha256": "de94d4b8e7ebe9b636862fbf296201cdd397cd3bde01c58953f6a297546e58b1",
     "size": 1775
    },
    "index": 47,
    "prompt_sha256": "fa0555b7c7c9b0aef964e7bcb0ee8c9551171c6e2e1aaf19c502a8aa99072765",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/d90b2f9ba6f52e30b6912c7a1dd71ea8ead3c04c1d9c3fc7a3445956b65aac73.json",
     "sha256": "d90b2f9ba6f52e30b6912c7a1dd71ea8ead3c04c1d9c3fc7a3445956b65aac73",
     "size": 5439
    },
    "index": 48,
    "prompt_sha256": "eb1c31b9cfc7d08696c9e211afb5615139f4fd764932a56eedf650f23e52eb74",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/bec53e7cf067b1db21eb4205a524ebde29c31051c83011d9a2ee114504fc8f82.json",
     "sha256": "bec53e7cf067b1db21eb4205a524ebde29c31051c83011d9a2ee114504fc8f82",
     "size": 4644
    },
    "index": 49,
    "prompt_sha256": "70b8e601417097e568f607f150d51bf4324ddcaa945ecf23795153e75ea7c097",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/666e6bf6d42b2862db5e09b25064e1936413492b54286c7bc7947dfb6e7e248b.json",
     "sha256": "666e6bf6d42b2862db5e09b25064e1936413492b54286c7bc7947dfb6e7e248b",
     "size": 4875
    },
    "index": 50,
    "prompt_sha256": "e45f0a236bda8ef8a8037b3acc1d2f9490e8794d24bd84ca0f365ef71755d353",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/75dbfb3f512c37b33877538200790dc217794ef2312b0c8cea35b5d3c2459f31.json",
     "sha256": "75dbfb3f512c37b33877538200790dc217794ef2312b0c8cea35b5d3c2459f31",
     "size": 5213
    },
    "index": 51,
    "prompt_sha256": "d5dc138f69d01d6236b64c2c7edb5c43f43fa04a620c0ecdf6af6111035f6613",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/5047d552580f4414c98d5a625f539c49d3b46abfdb92b9cd1b02516c94547226.json",
     "sha256": "5047d552580f4414c98d5a625f539c49d3b46abfdb92b9cd1b02516c94547226",
     "size": 2814
    },
    "index": 52,
    "prompt_sha256": "d4a5fa0ee4727d61abaf6eadfd0e0b5544509ff6818fc48cc7be093b3bcbd08e",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/b667d88463ed772de060153eae7f6ae68c68ef02d78837d104d560a0413d271f.json",
     "sha256": "b667d88463ed772de060153eae7f6ae68c68ef02d78837d104d560a0413d271f",
     "size": 2018
    },
    "index": 53,
    "prompt_sha256": "af83b5896482e8813ff0ad75e5936b2df2fd0ccefb2ce632ab5996cf38a13fea",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/d69d095bed5b10568c7a8a7e14cf1550db6f394eadc0097d2264ee27567f2aa1.json",
     "sha256": "d69d095bed5b10568c7a8a7e14cf1550db6f394eadc0097d2264ee27567f2aa1",
     "size": 2206
    },
    "index": 54,
    "prompt_sha256": "9350c97801ff8f4a9d41504bee3589f1a722878d37216024f84e9c96749cb83b",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/f60357073e6511dde7e4d2df1cf7c37441f731ce1809ff4c2ece6330b41f017e.json",
     "sha256": "f60357073e6511dde7e4d2df1cf7c37441f731ce1809ff4c2ece6330b41f017e",
     "size": 2264
    },
    "index": 55,
    "prompt_sha256": "e5af92842d75d9ba15d786b450934ef21920d50053f32525dc4f248565e4358f",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/282b3996e53679db88bdb8a51e43692b0a3274b33208020bf140e708397a644e.json",
     "sha256": "282b3996e53679db88bdb8a51e43692b0a3274b33208020bf140e708397a644e",
     "size": 1814
    },
    "index": 56,
    "prompt_sha256": "172e0f2223f305fb32244b321e45c2f51fd17fbc56a8cced92a38fd9f38d0892",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/a46021bb5db33151f14ca7e114284e0a3da906214d99260f2a83b8bb13e8e2d2.json",
     "sha256": "a46021bb5db33151f14ca7e114284e0a3da906214d99260f2a83b8bb13e8e2d2",
     "size": 2538
    },
    "index": 57,
    "prompt_sha256": "45717fb8199bf313771ef4ad572a1d2753ce8d51836ed603d2df5a5ae6dd7216",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/dfbd929535e10a54df8134dc8eac2e5f9cc99f51c4cdb4434a1cc6641ad63e4c.json",
     "sha256": "dfbd929535e10a54df8134dc8eac2e5f9cc99f51c4cdb4434a1cc6641ad63e4c",
     "size": 11454
    },
    "index": 58,
    "prompt_sha256": "889def1b34db7ba6d7bf10c119f74daad822f2c67d8bdb6b423bf54099864e6b",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/03324f3aba73c04352ead2c4c1c6d45b78145a33dab47f1266f5dfcd05b8ac00.json",
     "sha256": "03324f3aba73c04352ead2c4c1c6d45b78145a33dab47f1266f5dfcd05b8ac00",
     "size": 5794
    },
    "index": 59,
    "prompt_sha256": "508e1455ddadc2780350874b84db4b0bac692cbd5aeba779e9d77a921022613d",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/4d7c49401ff8521a5fa517d8183bbea492a5ccf974717dd7cac59a05fca54a72.json",
     "sha256": "4d7c49401ff8521a5fa517d8183bbea492a5ccf974717dd7cac59a05fca54a72",
     "size": 21544
    },
    "index": 60,
    "prompt_sha256": "4e7bbb42986affb4afe8f02a1efa1469af9b4a77b8935a25ed02a9630c2a83e2",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/5346224d9ee84c539690c5ef2d19c0d67acb4e5639b1a25f3a955ebf0716da3e.json",
     "sha256": "5346224d9ee84c539690c5ef2d19c0d67acb4e5639b1a25f3a955ebf0716da3e",
     "size": 1724
    },
    "index": 61,
    "prompt_sha256": "0d6b921e91d0868e2b5e4ddea9d57820255bda17d67c922d5bea8679f41bde31",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/ac3d0fca99a9e68b0e34e59568c42a066e7b1bd738330cd3fc9a563d048db54d.json",
     "sha256": "ac3d0fca99a9e68b0e34e59568c42a066e7b1bd738330cd3fc9a563d048db54d",
     "size": 2918
    },
    "index": 62,
    "prompt_sha256": "8c3e792c5de6b764c31c5134f568895d96ae3c0532b22bc649e7c44b0792ffce",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/85f4049a107d7772bd5af3d2809034be6e5cdb45f577efdfaaae1b6b3f7b2696.json",
     "sha256": "85f4049a107d7772bd5af3d2809034be6e5cdb45f577efdfaaae1b6b3f7b2696",
     "size": 3402
    },
    "index": 63,
    "prompt_sha256": "4dd47cc343d7bb32617c5beb73a7ee52a96570113327d4c2cf474cd7db581851",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/cc94fdaad8258c84e3efce1835ac42f52338a0c14d3806930d9861c1a6fbdad2.json",
     "sha256": "cc94fdaad8258c84e3efce1835ac42f52338a0c14d3806930d9861c1a6fbdad2",
     "size": 2245
    },
    "index": 64,
    "prompt_sha256": "f030fe4c7ba4297504efed86345b897235547563557d0a06c6b3374a4034d82c",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/cd469656b21f9cd90cca25f1642c2315a978882a173683ad2cebccf4b4433fc3.json",
     "sha256": "cd469656b21f9cd90cca25f1642c2315a978882a173683ad2cebccf4b4433fc3",
     "size": 1860
    },
    "index": 65,
    "prompt_sha256": "effa3416d476e16317ef59941a6ca6b7289fd53d5a37ee4b4035c67065c3fa5a",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/3cbbc002fdc9fcbef0af3094477730ed54866f45d4b483964b6f69cce80affa9.json",
     "sha256": "3cbbc002fdc9fcbef0af3094477730ed54866f45d4b483964b6f69cce80affa9",
     "size": 10990
    },
    "index": 66,
    "prompt_sha256": "56d4f8ead80055af1e1dc831d8c5d24564adb3b7d879688217e34ba582fde080",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/58b3f9489222d857179b09489f06b42edc611d719a328fa35fc03482bcd48932.json",
     "sha256": "58b3f9489222d857179b09489f06b42edc611d719a328fa35fc03482bcd48932",
     "size": 2049
    },
    "index": 67,
    "prompt_sha256": "117c645560f4ebfc4153899abb6014613605532f43484893da19462509a0f527",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/8a32d8598d66c7d3c60ba4db40844c786a2a7caccf8f039473521a09ead93cda.json",
     "sha256": "8a32d8598d66c7d3c60ba4db40844c786a2a7caccf8f039473521a09ead93cda",
     "size": 3543
    },
    "index": 68,
    "prompt_sha256": "bde19c0f3fe19d2634ffef86ccd72897d4590f8f0da81860db64574525bb526f",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/f13de95e78f732eb491eb12d3f49cb0746febbdfe43e7b346c8b77de95a4be76.json",
     "sha256": "f13de95e78f732eb491eb12d3f49cb0746febbdfe43e7b346c8b77de95a4be76",
     "size": 2991
    },
    "index": 69,
    "prompt_sha256": "12ee6b1d4374a1ea00f470fdc2b3bb6e2942fbd87facc6fc707ab9e2f624242b",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/89c8a8b9d04bc6587e8da98c3122a9e359b3a7d253db971bfb5c5f74a763035a.json",
     "sha256": "89c8a8b9d04bc6587e8da98c3122a9e359b3a7d253db971bfb5c5f74a763035a",
     "size": 2291
    },
    "index": 70,
    "prompt_sha256": "29e3d5dbbc93cfd7a637a8d926a791424837d938cb7761994fcc6ed66c1950a8",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/a597bd137acaede345419c26d79639c3943bd41010dae24a1eff690f5c6d14ab.json",
     "sha256": "a597bd137acaede345419c26d79639c3943bd41010dae24a1eff690f5c6d14ab",
     "size": 1013
    },
    "index": 71,
    "prompt_sha256": "421b70325c6b64f17c7fb1d4ee82b74335c65089ccda91eae7b5ed798cff184b",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/55218f372c5c1ef5bac8422e30d9bdf61d2d50ccf5342e3136c84c91497a765d.json",
     "sha256": "55218f372c5c1ef5bac8422e30d9bdf61d2d50ccf5342e3136c84c91497a765d",
     "size": 2541
    },
    "index": 72,
    "prompt_sha256": "259f396499f4688f71cb84b9885d94d7c6484cedcc699351429bb46ce3927e41",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/477fce3d60fb5fbb3a87dd8b28f98ef6b85bd7707ed81fb60a826c208d3effba.json",
     "sha256": "477fce3d60fb5fbb3a87dd8b28f98ef6b85bd7707ed81fb60a826c208d3effba",
     "size": 7569
    },
    "index": 73,
    "prompt_sha256": "798d21fe7c614e5afe3f8fe2212f1820d5d20d3382cf9c0b8437077a488a1b14",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/4f6505a06e8824e3ef20dadd730552b1d57708f469177a32098cb9a9a23538c5.json",
     "sha256": "4f6505a06e8824e3ef20dadd730552b1d57708f469177a32098cb9a9a23538c5",
     "size": 14955
    },
    "index": 74,
    "prompt_sha256": "c742872346efe5dcdeeac440b06d96faabf541e57019ce7989cb06319c7f227e",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/21ed1c6f8eba87255fc2d132434c7ffa44dd3594cf4d2eab077ccc171654eb96.json",
     "sha256": "21ed1c6f8eba87255fc2d132434c7ffa44dd3594cf4d2eab077ccc171654eb96",
     "size": 3298
    },
    "index": 75,
    "prompt_sha256": "a5070e595248316097d74f5633fd9659e21aa572451e8a0e5df50f8494092b70",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/1129f2290e2e41752ce0323a3bd6768ba72c0304e70f4cf18a8727e5219ad325.json",
     "sha256": "1129f2290e2e41752ce0323a3bd6768ba72c0304e70f4cf18a8727e5219ad325",
     "size": 13495
    },
    "index": 76,
    "prompt_sha256": "5cb53d873ee4d71d3df490831204943337a8219d6929f27b582be8af052e6382",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/59c400f4626d561c716a47eb44736bbe7cecf6514748b0fb25eb0bf36a5daea4.json",
     "sha256": "59c400f4626d561c716a47eb44736bbe7cecf6514748b0fb25eb0bf36a5daea4",
     "size": 5753
    },
    "index": 77,
    "prompt_sha256": "4f5321bd99a2061d521965a18b311466f6a0b24eda143c18eac00423c2c1a028",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/0006c1f465edd26e285b4c04a7b47154f98c3bb08e33310fc282f46ffdf231cd.json",
     "sha256": "0006c1f465edd26e285b4c04a7b47154f98c3bb08e33310fc282f46ffdf231cd",
     "size": 3788
    },
    "index": 78,
    "prompt_sha256": "412e6b2c340112d22e689b2ff9eb9b399b5968a31021817eb690a3a2765be567",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/a33371a2421293383d69a6a136caf04d1c2651ac316d6a7fe16d20c2eff9de56.json",
     "sha256": "a33371a2421293383d69a6a136caf04d1c2651ac316d6a7fe16d20c2eff9de56",
     "size": 4572
    },
    "index": 79,
    "prompt_sha256": "de6e9b9c52f7810302edb3b2d843bf8226d47a2d3cdea038045489dd88844d4e",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/38f729352fcbb3e0e8f4f5974f61e9a923d68f2a2d530d3409fd06de41912f60.json",
     "sha256": "38f729352fcbb3e0e8f4f5974f61e9a923d68f2a2d530d3409fd06de41912f60",
     "size": 1804
    },
    "index": 80,
    "prompt_sha256": "26f0c989909bb239241da5f742e892cf22e9090a53de01c35a13a62cdeedf9cf",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/5179b79850756da04903a80bf15a643b622066486fae5159ee009a93be943d15.json",
     "sha256": "5179b79850756da04903a80bf15a643b622066486fae5159ee009a93be943d15",
     "size": 7899
    },
    "index": 81,
    "prompt_sha256": "65e99a79c94c281ca30a4d2d604aec5faf445af19297a8d1bbcb7b0237aa89cd",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/98a7048a5c631c8156ef9b5326b034aee19d08aab5763ac0ef5f6f84149a79a4.json",
     "sha256": "98a7048a5c631c8156ef9b5326b034aee19d08aab5763ac0ef5f6f84149a79a4",
     "size": 10724
    },
    "index": 82,
    "prompt_sha256": "b7fe9ea0dc9a5b527695f7164a020f0e71acf9feda03b27543b6b6ce2f7c25ac",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/43b7e6db18b2e693905f0d020bfb7ffff68bf12ff4f0e682a672294396b5fcf7.json",
     "sha256": "43b7e6db18b2e693905f0d020bfb7ffff68bf12ff4f0e682a672294396b5fcf7",
     "size": 2907
    },
    "index": 83,
    "prompt_sha256": "295d98f4b577ddfbf23e2f7c0cd041f0c509b16dd4d3749576cddba26a0ae3c5",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/c66ded5a9d1ec643c0e227ffaa5c15c52172866e33b5ec484726e0dc88796999.json",
     "sha256": "c66ded5a9d1ec643c0e227ffaa5c15c52172866e33b5ec484726e0dc88796999",
     "size": 2345
    },
    "index": 84,
    "prompt_sha256": "9e38498427b6f5758c5652df9fc1c7770b47a4d8f7d188281f874a8a9d6ec82d",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/d9c5f48adc25f8a92f785625310088ef7603ceab321de00d63131b51a1c702de.json",
     "sha256": "d9c5f48adc25f8a92f785625310088ef7603ceab321de00d63131b51a1c702de",
     "size": 9652
    },
    "index": 85,
    "prompt_sha256": "21396a9a616ca483f009f69323d82099a1b8ebdac2cc2d20dcb8f13b4910374f",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/82afaceec3b7cea70115addd8d74d50dc0b5d094236c93d776ee8091abf63522.json",
     "sha256": "82afaceec3b7cea70115addd8d74d50dc0b5d094236c93d776ee8091abf63522",
     "size": 2954
    },
    "index": 86,
    "prompt_sha256": "7e28f36f871b73481813b583c56e606e4aa8c98e0700288689d8e66950f78243",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/d15b4028d6744bd7ad7cb7696caf5c5084837ea969d41d317fd0074085ac7433.json",
     "sha256": "d15b4028d6744bd7ad7cb7696caf5c5084837ea969d41d317fd0074085ac7433",
     "size": 6390
    },
    "index": 87,
    "prompt_sha256": "a175d16d272acce217cb719bd976c099652dad17faac94883247b991d51a986c",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/bf50fdb67002c17ee30ae6683f7366622d139d972d603e9ba30e9a661009eda0.json",
     "sha256": "bf50fdb67002c17ee30ae6683f7366622d139d972d603e9ba30e9a661009eda0",
     "size": 7469
    },
    "index": 88,
    "prompt_sha256": "8f62abe5534080d53df8db49f1eea7fa71df763da2da245b0a8e213b619dacb4",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/cc25d530fb93b42243e51869d24c42807c59465f14ce9f8e0e95888e2c11f7a5.json",
     "sha256": "cc25d530fb93b42243e51869d24c42807c59465f14ce9f8e0e95888e2c11f7a5",
     "size": 3163
    },
    "index": 89,
    "prompt_sha256": "414e199a8d8de86c39a9e98038e861f93ec1930573c1f639eec9ddb33981847b",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/240425d3b39be9d2462aa2d8f796f51f3a41d8020948feffa93dccbccceface4.json",
     "sha256": "240425d3b39be9d2462aa2d8f796f51f3a41d8020948feffa93dccbccceface4",
     "size": 3243
    },
    "index": 90,
    "prompt_sha256": "9b0980dbc93ea03aa16e74e355ff00c6722855849f97ce218dd19ffc2ca5b926",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/456ddc1f6ff3fee23174037db54ef184caf731802104f05d01c30e5e886e3218.json",
     "sha256": "456ddc1f6ff3fee23174037db54ef184caf731802104f05d01c30e5e886e3218",
     "size": 10170
    },
    "index": 91,
    "prompt_sha256": "415cef54cec36bc25ea931ecee33892d3174dbdac6d69a3e4c34212ef2757220",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/9c2633293500458e2b730a875a0ac7507195d870494b0bea6d6430256dbe5da0.json",
     "sha256": "9c2633293500458e2b730a875a0ac7507195d870494b0bea6d6430256dbe5da0",
     "size": 1452
    },
    "index": 92,
    "prompt_sha256": "3a8407d1d18a922ce328d128c0f97192cac2f5efecc43e10eccbd0bfd46305a1",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/d830158fa0950fa75bc615912fcea27efe13c227e2a20db3c7d309faab742201.json",
     "sha256": "d830158fa0950fa75bc615912fcea27efe13c227e2a20db3c7d309faab742201",
     "size": 5025
    },
    "index": 93,
    "prompt_sha256": "40d018b154fe58eeda6b0a461bb0a601f0b1fbed89c5d7be90a71c4bd373d05f",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/7de7828a4d86bc8d2d095a0af3ee9185d87168cb119058267351e2b5106596ae.json",
     "sha256": "7de7828a4d86bc8d2d095a0af3ee9185d87168cb119058267351e2b5106596ae",
     "size": 2786
    },
    "index": 94,
    "prompt_sha256": "f5fc5cfadedd6e953ba3c966e7d93c40b180cad0e2e5db006a31c422fba9f427",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/7c08e6a9976e1f6c8c91f91b694b1dd2a6a40b233bf4542199194dd826d69ad1.json",
     "sha256": "7c08e6a9976e1f6c8c91f91b694b1dd2a6a40b233bf4542199194dd826d69ad1",
     "size": 13875
    },
    "index": 95,
    "prompt_sha256": "ddd5f35adae44b9270d0d5bb7fc4dbc3e6dac53c9ec8b49094fb36261820a70f",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/a012aff552f3745fb739c882620c5177500193a82b6b4b6bba75ddae693a66b3.json",
     "sha256": "a012aff552f3745fb739c882620c5177500193a82b6b4b6bba75ddae693a66b3",
     "size": 5651
    },
    "index": 96,
    "prompt_sha256": "423c800a55fd0a7c39977b2ffa4cd525746895bb4ac42b2a3ca4dd998cd97751",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/1bcfda598caa33fde55003792e61dcff00f3faeca645eb47037908512b1f065f.json",
     "sha256": "1bcfda598caa33fde55003792e61dcff00f3faeca645eb47037908512b1f065f",
     "size": 2362
    },
    "index": 97,
    "prompt_sha256": "d7ad5b32da03ae1582c1ea360f250e4da07d192105114d23b91910878dc8cde5",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/2ac778307ff8be909b854d6a2ec348c6b0e6e4071c218be9bb908ced93034159.json",
     "sha256": "2ac778307ff8be909b854d6a2ec348c6b0e6e4071c218be9bb908ced93034159",
     "size": 4218
    },
    "index": 98,
    "prompt_sha256": "a6e282459e7b0f8adb16a1a509da5b0f91c3f3c3e8930b36fd46604e848c653e",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/906dfe9f21e2dccd2d52ee33429f146e356598e7f56ce37836135c54638ccfdb.json",
     "sha256": "906dfe9f21e2dccd2d52ee33429f146e356598e7f56ce37836135c54638ccfdb",
     "size": 6575
    },
    "index": 99,
    "prompt_sha256": "c556b334d0033518e7774681c7576c019b91b53ea7eebda412188cd0de7ca8f1",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/e8e32843d4aa273f2b4df5b83212fd51ff099724852b382df80d174b4aa6a3d2.json",
     "sha256": "e8e32843d4aa273f2b4df5b83212fd51ff099724852b382df80d174b4aa6a3d2",
     "size": 2760
    },
    "index": 100,
    "prompt_sha256": "efde1c494fe92e8206a98b5d841f6e656aafa2dd8465ad13b4fad9be9d028562",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/97eb2611ec1a5a67a04ecdb4f8a6ca72fbcec5dda83f44bd02437e6e8e23d88f.json",
     "sha256": "97eb2611ec1a5a67a04ecdb4f8a6ca72fbcec5dda83f44bd02437e6e8e23d88f",
     "size": 16059
    },
    "index": 101,
    "prompt_sha256": "22458f6ffa018c1ddfb28ba7f17daebe3d282421a2889a25155ce369e641f0e9",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/00921e7e1b2a69702ed5f4ebd3e4f3f1974af8d37c60a02a27057df4498ce891.json",
     "sha256": "00921e7e1b2a69702ed5f4ebd3e4f3f1974af8d37c60a02a27057df4498ce891",
     "size": 6546
    },
    "index": 102,
    "prompt_sha256": "b1b7ed6953310113b27dfb97374501fe8675dbff3396cfaf2cc48f2d0b293046",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/4aa9c68d123846cf4e3d01787957f16e68791e8ca6a5526659b13ad60ca87cf6.json",
     "sha256": "4aa9c68d123846cf4e3d01787957f16e68791e8ca6a5526659b13ad60ca87cf6",
     "size": 5699
    },
    "index": 103,
    "prompt_sha256": "1a4d1f5aff89af03ecd2cbd8df24611345a087bbe515d96071c78bbf2648e69d",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/178dd7ebb5a6194d058fa772ceeaca2eae0c0d4ff3968c3cdd82bd5245108e27.json",
     "sha256": "178dd7ebb5a6194d058fa772ceeaca2eae0c0d4ff3968c3cdd82bd5245108e27",
     "size": 2164
    },
    "index": 104,
    "prompt_sha256": "832a820fa0f5d124bd0fb352761fe5d52fab2717d6447d5b7b2ac6054dd28776",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/6a526b6211e1f9f2554089e716c8e928857624df52eeba6d11c297b631737ac7.json",
     "sha256": "6a526b6211e1f9f2554089e716c8e928857624df52eeba6d11c297b631737ac7",
     "size": 1984
    },
    "index": 105,
    "prompt_sha256": "5e5ce39e3bf7644b3f3d5bae5581a9ea9a882cdcd35c7c78f3fcc528b16f95ca",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/741e40808469bf25c451805863f2cd94317a31868079e963e986471eed91ccf2.json",
     "sha256": "741e40808469bf25c451805863f2cd94317a31868079e963e986471eed91ccf2",
     "size": 2417
    },
    "index": 106,
    "prompt_sha256": "d80f2e7bdad3e3661c0d68f0458419a2b2372aa946d11f7a9760522ca171ef0e",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/540c4f428666e55c0d5c3d670067a5baa909b0ce4c1649f888744d70a46d22df.json",
     "sha256": "540c4f428666e55c0d5c3d670067a5baa909b0ce4c1649f888744d70a46d22df",
     "size": 1591
    },
    "index": 107,
    "prompt_sha256": "febd85a764ac5ef5927f138726925b82191b632a4d7446776fdf125928c1e5b2",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/2fab73b78c74c8e80a3e5a16db89d78c3233702b900887b8059054a289a2cd57.json",
     "sha256": "2fab73b78c74c8e80a3e5a16db89d78c3233702b900887b8059054a289a2cd57",
     "size": 4414
    },
    "index": 108,
    "prompt_sha256": "3dc17b5243984f7a50885617e9e724bfc656bfaf0f5a218f9765ef51944d258c",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/7e447f3a20c1d6ddafdf9669888f39ff2a87307e525dd6d46489743baf0d52f2.json",
     "sha256": "7e447f3a20c1d6ddafdf9669888f39ff2a87307e525dd6d46489743baf0d52f2",
     "size": 4125
    },
    "index": 109,
    "prompt_sha256": "268869ad82f86635672a1100fca457134a48ea7308046808bc6c67c0a66d9f8a",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/65b618e611183a7ae06a2de3633eec0c9abfcebc4b83a708c2d568c11e82bbf5.json",
     "sha256": "65b618e611183a7ae06a2de3633eec0c9abfcebc4b83a708c2d568c11e82bbf5",
     "size": 6332
    },
    "index": 110,
    "prompt_sha256": "0cf501164dd202d0817e99aea523439cc851aa5717c77b2040cb8c8947d5d112",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/a4be1b483b79af626bee7c0883c5eb00430e502ba66ed3699e04c948376f4310.json",
     "sha256": "a4be1b483b79af626bee7c0883c5eb00430e502ba66ed3699e04c948376f4310",
     "size": 4307
    },
    "index": 111,
    "prompt_sha256": "a6972574328183e1d3e328ff090ccbd3019821692aafa97ec40b5f2ac337ff62",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/3ebe73c91b290991111df8b905abc505da8adadb02e3d6bce8947061b7571db1.json",
     "sha256": "3ebe73c91b290991111df8b905abc505da8adadb02e3d6bce8947061b7571db1",
     "size": 7810
    },
    "index": 112,
    "prompt_sha256": "00bd56ffecc55d1f0660887a21e455f50437f9f4079c76cd9d36b7184b94000e",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/ad521a4448b7e47a65237fecc651dd1e03c9970f33f48e992fcca048250ffe40.json",
     "sha256": "ad521a4448b7e47a65237fecc651dd1e03c9970f33f48e992fcca048250ffe40",
     "size": 16534
    },
    "index": 113,
    "prompt_sha256": "444ad7f711e3529374e71c62021d1900dc72f59b1c77428abf26dc3bd971a8ed",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/1917d2e20e5932f4dd8bd786ac99cff734556f15e0a888ab30332dd2947eb0ab.json",
     "sha256": "1917d2e20e5932f4dd8bd786ac99cff734556f15e0a888ab30332dd2947eb0ab",
     "size": 1391
    },
    "index": 114,
    "prompt_sha256": "031dd048dc7591d85d33f438aa511dc6d141a390486e174f2b439179f54851d9",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/0ec8e7f7aade1c782087b16a70aa487f791d81e6ba8bfac57367c6bd992193ce.json",
     "sha256": "0ec8e7f7aade1c782087b16a70aa487f791d81e6ba8bfac57367c6bd992193ce",
     "size": 4026
    },
    "index": 115,
    "prompt_sha256": "8306d4805f730eae91c1e1f553bed8da174585f0ffeffe6c9012f427277f0c44",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/16cde55391c71ddfa87fcd06506aef819a42481fd5a7127e3f7bd341a48641ad.json",
     "sha256": "16cde55391c71ddfa87fcd06506aef819a42481fd5a7127e3f7bd341a48641ad",
     "size": 2074
    },
    "index": 116,
    "prompt_sha256": "821fd177ece0c4de638973224075eddb3de9fbfb537662141363213c24c9364d",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/1224ff3a86d8d8b011b812468cbae50185bcc536889a4f842063c36a8378643b.json",
     "sha256": "1224ff3a86d8d8b011b812468cbae50185bcc536889a4f842063c36a8378643b",
     "size": 21996
    },
    "index": 117,
    "prompt_sha256": "e1c6be63bc5f0ef31e3435a205f9a73f3784304e00fcd997e4de302b1cc0381c",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/8e42f8ad0388b58c4de7b1d4855175edec7d474ebbf7a4a1d54141f79595aaed.json",
     "sha256": "8e42f8ad0388b58c4de7b1d4855175edec7d474ebbf7a4a1d54141f79595aaed",
     "size": 2351
    },
    "index": 118,
    "prompt_sha256": "4adea13a8e4d47b7aecc4f94a4356bd111efe24cf0be7b8282228910aa1b038d",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/c9a587e7c5805ca5469d07d21e77d5b8a48952835b28fa5e5444c9c03c248bea.json",
     "sha256": "c9a587e7c5805ca5469d07d21e77d5b8a48952835b28fa5e5444c9c03c248bea",
     "size": 2404
    },
    "index": 119,
    "prompt_sha256": "d6a3d7dffdb959aeb891ed9f39a227cbafc2e66a6aaab31b934f4a4699e341fd",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/a5887e186362d5375cec37765e694e80e97dfb30200dbec783baf2e40b512db8.json",
     "sha256": "a5887e186362d5375cec37765e694e80e97dfb30200dbec783baf2e40b512db8",
     "size": 20998
    },
    "index": 120,
    "prompt_sha256": "eae849774973dc6e18b0c16a2a2ffe3746d9ac584cab7349774d76892dca3507",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/1e53300af3e55af8c525e10ca101b0237fba7a718786eb5a6a8069d3bfb02452.json",
     "sha256": "1e53300af3e55af8c525e10ca101b0237fba7a718786eb5a6a8069d3bfb02452",
     "size": 3294
    },
    "index": 121,
    "prompt_sha256": "75dff9f1deab5ae07fcc7d0ee2e57ec0a6ec026a93f080a67bd9df87e4b7ad69",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/4f97fecd88f4d6f44d776bf426c7263f6769af1257866f103545f3dcc0b78590.json",
     "sha256": "4f97fecd88f4d6f44d776bf426c7263f6769af1257866f103545f3dcc0b78590",
     "size": 3861
    },
    "index": 122,
    "prompt_sha256": "f8868f2e516b55c4d99c55c68b74703e2e8a3a1d9631edbf19ac872289a88b1a",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/76ce8c4b89c15eba127643f7076804d2cb6e9f0275ff39dfdf67ec6bba1af417.json",
     "sha256": "76ce8c4b89c15eba127643f7076804d2cb6e9f0275ff39dfdf67ec6bba1af417",
     "size": 3179
    },
    "index": 123,
    "prompt_sha256": "d1a407d79a14a82fd47ba9bcb27768aeb73637a1f8e17fdaa104eb4f5c65b43c",
    "rollouts": 1
   },
   {
    "document": {
     "key": "private/ff/v1/turns/affine-mimo9b-grpo-20261013-e000354/docs/db1de19fc6571348e69516ae57dc3ab967a3c1243cd438e7609bbf3cd4585016.json",
     "sha256": "db1de19fc6571348e69516ae57dc3ab967a3c1243cd438e7609bbf3cd4585016",
     "size": 5271
    },
    "index": 124,
    "prompt_sha256": "f3483bfe36a3118df624c40bd415abed6a85cb8c586003f1f8d0f9d2dc386fdf",
    "rollouts": 1
   }
  ],
  "signature": "ks2ftC7jhsd3mzm2aLUuHrO12NNo8wnxrYlwmY8QE+BAxcoqlEmcUI6AJaqZcsSVDXlQJ/xV8WlHWus1QVMsBg==",
  "signer": "3301134b38401196d006a621ac4a772bb4b0e6afa15a7d34a0d1ae5f2c630bcd",
  "source": "lifecycle TRAIN-DISPATCH-FF (signed): one document per prompt group, rewards as given"
 },
 "engine": {
  "status": "unavailable: local trainer host config does not match the operator config hash in use"
 },
 "epoch": 354,
 "format": "affine-batch-v1",
 "gate_reveal": "/gate/reveal/354.json",
 "promoted": true,
 "published_at": 1793453968.0,
 "recipe": "/k1/spec/step-recipe.json",
 "runtime_versions": {
  "toploc": "0.1.6",
  "torch": "2.14.0",
  "transformers": "5.14.1"
 },
 "schema": "sn120-k1-transparency-v1",
 "training": {
  "admitted_groups": [
   {
    "admission_sha256": "f51727fb52fddbaecaafcc5112bb0812ce1dd550d3fd645ede58972af956be3f",
    "task_id": "47623a99cd3c84cc3f5b106c5117e6a0d4de2957aa7628ac468a7777bc3c8e9b"
   },
   {
    "admission_sha256": "cdae5c9b95d55580d21be24f7cec74cb11a6a93f6748723cfedde273a7d0f51f",
    "task_id": "febe060739338ff7b01afd1e01f1c31072fa1095f3b379a68b7ecbd6807790c6"
   },
   {
    "admission_sha256": "a2837fe3ca4f53360c7a1620d79c32fdffd043a311d26ff27ff020af75335be6",
    "task_id": "6e13d4931d8cb01e03dbdf22d0952a61aca999fa63c232632a32a50cccf2f47c"
   },
   {
    "admission_sha256": "62f0164603dbfa981995fd345b78383168713b56116e4c912750bc45ebe14fd7",
    "task_id": "0b33db1a5b7ab3ea5eaf737d908c78452d416b5682203a8ae92dd2b9aeb02630"
   },
   {
    "admission_sha256": "76574e4c6026d87c9ec8b5b0753af9ba98efc69e9b2c15f44b56718ac5b16a2b",
    "task_id": "da1e6903b4411949cdc2e27f2596bc1eb0a51e2477da503f3e243b2f9ad517ce"
   },
   {
    "admission_sha256": "7ed38c9c325200599e732327d4c258f475c2d7994dc36a61ba6e981f9e3cc3b4",
    "task_id": "5cc075ace2ca1500a2d3ac6f380af314007146cc7f7c083e5d8094ccc675c1a1"
   },
   {
    "admission_sha256": "d7d17185ca4e390aa57f16bb20a7648dbdd516ef542e747fa0f38157e314daf9",
    "task_id": "728572665e3b46f5bbc74a0bdae6725ff509683850a32d041863c715f8fb0c60"
   },
   {
    "admission_sha256": "51fa4058c6af3b4963e5bebcb7d2ec2ac6a87e762fc897684547f778e0dea68c",
    "task_id": "588fa076f9a7e0e5062c2d7fee98c01524c7798edea325f3f0f52956eaa9a35e"
   },
   {
    "admission_sha256": "e5010f901dc63185155af840a3fbbbe76fa4886ff323a07985bd402fcbe7a994",
    "task_id": "dc864ee0a497b0af693911a96d8e566d425de2aaa2bc3538522262123453d1a0"
   },
   {
    "admission_sha256": "831d0fc1352e110befbc1320aa4c13da85ec273ac7273eaac74bbc083684ae34",
    "task_id": "bdcec23f6afa3d95df26b3f33e880fe9ddd11713c8955e53ecf137445de108ea"
   },
   {
    "admission_sha256": "06ec1a059302c317c20e12199ae833ae73a7688dd3e260c835eb5e1b9814dafa",
    "task_id": "4e53d8ae085f72266bfdf020dcd1bf2b408ac3972e2e174b04f341e4f64f66e0"
   },
   {
    "admission_sha256": "66b5afa49d182228ecec3b774015577a5dc39a3207b8512d3bca877b5ca0b210",
    "task_id": "acf03f84995c27e667b9d698105ddaf4a184ecc0c6cdcba2c48b271c99164c8f"
   },
   {
    "admission_sha256": "cf330ee5ea9fabdb58d30926e710e8360aacc3d0406678f2ba5d937d45f1009a",
    "task_id": "c0b7ffc2401635b3c61b1b3d7099332f5d511bd5f46ea13920b88e39d678bc46"
   },
   {
    "admission_sha256": "4db533b755971570d6003fd749710d4e8b8cf1a189d3e173ec78a9993d30c8b7",
    "task_id": "211c35a89bece9052a56bf99ebc406ef800a63fc7c9992212a9685240a27353a"
   },
   {
    "admission_sha256": "e18d81d4364db6b1d8b797e4017b1e7116fc6b71eb800077cf20888c5f1170b9",
    "task_id": "0b407bb90ca657a02eeeefff97e3ca30ed492a5fd569d3e562bc8ef24ff5decc"
   },
   {
    "admission_sha256": "699d38b1d8ca02fc2f102a06cc72adfcad75343336e206acbd8f909577c06989",
    "task_id": "0545c87932bc03d0afccf9ed2ed02a227e35aa480d17c17a5c303b73943c2ebd"
   },
   {
    "admission_sha256": "43d3b210ab37f9412548d2e636a011ac8bb47ce9718dae61ce4bc61b40e7730d",
    "task_id": "3a9c9c2393285164d3ec1564e3b583bdcce77d2db293864744b3aa29e05ba3da"
   },
   {
    "admission_sha256": "de0c6486933e75c628f1ffb420b2dc3775534a74925bf32fd9d95e5d86a3e6bd",
    "task_id": "41c6041825e6613a8f26a91de7ccaa69f3e66197bdfe52b5c564b760459106dc"
   },
   {
    "admission_sha256": "7ff6c2b4d2f4eac98bad134ca7df7e9f81b0913333dd585d86358da42b8bd225",
    "task_id": "f49dcac52bfecdf4236177cec61cfeff17f0788bdf286e7d80c5893cb94fd043"
   },
   {
    "admission_sha256": "458263c23337ac96eb3912f51fa81bd6d44dbd4ca4cabf85d8aca89278675087",
    "task_id": "a5609a80782de7a379bcb404ff62841e3e9e33819bdb028f5ed3e1b96e7733af"
   },
   {
    "admission_sha256": "c22bb4abfa5e63256990935b9def714486b3e753ee5ea743848dda049c745681",
    "task_id": "31a27a7b19d9daf0248fa3d6cd2d69fe535ed8719ec9ce4d3874e637ed619363"
   },
   {
    "admission_sha256": "2f89b8f4fe084f6a08efd61132633472202dd4c56ac0b019cb2437ad9bc7b135",
    "task_id": "a5c1a99dea52352c5e56a5b9d9891d71dd94b051208fff78df104277a66c6706"
   },
   {
    "admission_sha256": "ecd627cf58e20d238250c299c72a71a67a49b40e80983a0c4026bf1351caae94",
    "task_id": "0bf10d5d4f671583dbc6c168a304b883db92684e8cfa7105fb9603766c332960"
   },
   {
    "admission_sha256": "027d24a8af02196273bd291df3660e23eb72fecaa5e897122a3706c1b68bdece",
    "task_id": "13157aaeda469fde94df9ea04628c7e129e8fe7a21733fa6f53beac4b30dcf88"
   },
   {
    "admission_sha256": "7652b3b6580e3900897dd234e60f4e1ac253e4f1cc41b75e30a209a0dace2ea4",
    "task_id": "7284c6fe04f9fcc35cd25bac713b5f127d011c21be703cdeeb9a6e590ff5384a"
   },
   {
    "admission_sha256": "09527862bd9b6de0b04ce20d18115df2a01c04fea532b05ccb88e729545cbef7",
    "task_id": "76134b1a83640bb29536a3a0184943a5c42483780b8ead987c29c59e6e371ee1"
   },
   {
    "admission_sha256": "c9e4ccf0db9641053a3ccc16bad8c89f939d6dcfabf1ba90a7445ae3f8211442",
    "task_id": "d35718c793470eeb8f2c61109b019cba75bb70c623b284fcc1cc48351570819f"
   },
   {
    "admission_sha256": "71465c2fc02d7572756c675e8af1ab3f1fe588f74fbde72daee7ea57cc7eb3d3",
    "task_id": "ef9fc35202da7ab179894c4fc6253d848166e49461d2ca20525714274ebb90ca"
   },
   {
    "admission_sha256": "e2bd9e706778864e19ddb8e29ff6d4d487f6ac55d0d3d1241c42b133fea21dd6",
    "task_id": "9fd5af81c675fe33c89e89e8d235e9fcd7aff2775f881558d262a02fbbfbedb5"
   },
   {
    "admission_sha256": "199036a122b41e4f5ac57593a63272e70b3a81fea4a2ff7b1d87d352c511ce12",
    "task_id": "a607568da5e76f24b85d930f0b6ade1b4b0ca72c2b73c5845492699b9cbb48ed"
   },
   {
    "admission_sha256": "b02b2260dd2f4fdbd79f02beda4874bd27591368d1ab99b88c876ae43badacb3",
    "task_id": "58492be693417fc6b46aa5a8f58e8ba385359e7362d029145fdeea0f859fb14d"
   },
   {
    "admission_sha256": "dee6297b0a38f760b6e7e35cc1891eace2f32ac83802bfdff68105b65176069f",
    "task_id": "9bb2c40065b39cb6ba35bd202b8f40b821241e5e318d6478038b743b398401d4"
   },
   {
    "admission_sha256": "15b768ff9ecfde5eabaa16431599c876ae94e0a14389ba40a06b11d810323261",
    "task_id": "5c1e708d228de272a72281f5d2d00f37bf4573000ab8fecff240d1c4316127e9"
   },
   {
    "admission_sha256": "d94e1c383417745e9a8e0e851ba906b2fd05945499f8ff6885fcfda89de9f8cc",
    "task_id": "2e497d1c665174f37a39c52e1bc2b4910930439654c7f13cbeac8a508b106a78"
   },
   {
    "admission_sha256": "025d1a3fcab927c5adc7a20970593f51c3ef4e0f0feb8aebda16e18bc1614afc",
    "task_id": "68692eaa82eff5079d33d75e4a4ef06adcf3c17af4d2419ac46952046cbfd52e"
   },
   {
    "admission_sha256": "b8216db70db97bf84e244a32983243a9021e5e7c6f3cb3aea94ef999788b238c",
    "task_id": "483d8493e9a1357024ab8fb67fe1b8e54171285fb53d844ba4a17f7d795733ed"
   },
   {
    "admission_sha256": "4bf0c80c2e014736129058d85ac6d4b6e22c03041504d47aebfd81db0a76f597",
    "task_id": "8b321da461edefde469931f945899bea6bca0fb97c14becb05ed742a96ec5932"
   },
   {
    "admission_sha256": "255d826bf4a55b26c39ae7e2b261c7e25796fde9f69112c33ad6cc6d123134e9",
    "task_id": "fff4b30422ee0d3f21ef8d031119b89801ac5040797fa7637d11189ae508e9c3"
   },
   {
    "admission_sha256": "88e6bc2095ad4f359c5bde1a06b5911bf2439e1b5b262c3d8115a34adcade402",
    "task_id": "aff3cebcdd471ce6cc3923f46651c3e60a1c60a659e3ee6152c03ef0db6574df"
   },
   {
    "admission_sha256": "6a023a6665b4427afe118b5ca045bf59b0b884e19cea4af8d8e2d01a8fee1284",
    "task_id": "d71d5be03fa429509ba448dbd9db6d8ec2a77646c68fc142efd61ef4b51703f3"
   },
   {
    "admission_sha256": "8735aeee1d1dabe83798a67805c1c1b74c5a6d7f0d8378b996a536a2a6e1595c",
    "task_id": "cc81149f4ae6c4a0d8e8276388b86b3ef48cc96a194f9cbfa456c6e62b9018dc"
   },
   {
    "admission_sha256": "7b8825e53408016a1f01a3abbf625a110e3cedffaafecbeaf4857c7123a47d5e",
    "task_id": "cccbcd368f04fd6849292d3edbd9498b41702a38ef0c44e35ccd386e96017769"
   },
   {
    "admission_sha256": "1bbfe6a909ce1700a9ae065d5c28302b89675673e6aa8d23c032fb7a48c3a43f",
    "task_id": "36de2ba9c7a80873493f46295f9716a0f84ab342fb9de23b5c0256cceb413969"
   },
   {
    "admission_sha256": "906cb7e3326f6b468f34ee3cf4099495c363461d3fb4d939b6700e73641bbcfb",
    "task_id": "6012373932147e4abaae370d3d34472bf7b49177d95826a4687b3fd1985921c5"
   },
   {
    "admission_sha256": "87535617e445332e7bc91c6e457e854a84f034cdb0b2e31ff21a90b6029b9b25",
    "task_id": "8ce404030fc292408b1610d7dc822c1da42436a17c30d52042af714726b2bf5d"
   },
   {
    "admission_sha256": "5a401a102f9fe80851861a217ca4df71de33fcc7daa4566cbeeff9015960641e",
    "task_id": "c8d0735cf958b0914d1b0f795116684dde65ed0f73f5236bc2db83e42ce16129"
   },
   {
    "admission_sha256": "9969b33784ea2bb3da21a78a45a3a0d71e7b8ddabb8c00701ccbaab3306ec2fb",
    "task_id": "86cfb1c502c1d7f11de521391a9244c96ecc09c13b218a0063a7fb32650eaa22"
   },
   {
    "admission_sha256": "de94d4b8e7ebe9b636862fbf296201cdd397cd3bde01c58953f6a297546e58b1",
    "task_id": "24c8a25321edce466f3e3f5b40fe7e00739a039d434220810ae5a8d40ace4589"
   },
   {
    "admission_sha256": "d90b2f9ba6f52e30b6912c7a1dd71ea8ead3c04c1d9c3fc7a3445956b65aac73",
    "task_id": "e87b6c1304d795f10e08fb95c7eb23f2f9aaea6c268fd81e339a0bef1479e2b0"
   },
   {
    "admission_sha256": "bec53e7cf067b1db21eb4205a524ebde29c31051c83011d9a2ee114504fc8f82",
    "task_id": "ba6d93ec1099be804f5fbbd88423df4806d321248eb7c5920405f1b18d29262b"
   },
   {
    "admission_sha256": "666e6bf6d42b2862db5e09b25064e1936413492b54286c7bc7947dfb6e7e248b",
    "task_id": "293a02af5e505e4417db029c96706265de72dde0246b86b6e854cf36451b28f6"
   },
   {
    "admission_sha256": "75dbfb3f512c37b33877538200790dc217794ef2312b0c8cea35b5d3c2459f31",
    "task_id": "4196968a801934f0e876666b164e7afa49222d8f948d2921de1e72d9723835b1"
   },
   {
    "admission_sha256": "5047d552580f4414c98d5a625f539c49d3b46abfdb92b9cd1b02516c94547226",
    "task_id": "aedf1c445322c33c1fec5b93c17f36b951b0d65ebcf95d38a53f669d0ffecf96"
   },
   {
    "admission_sha256": "b667d88463ed772de060153eae7f6ae68c68ef02d78837d104d560a0413d271f",
    "task_id": "ccf7c349ebd12019071e6d746f0e981fb57b52b15b130c259057da060d13a0f9"
   },
   {
    "admission_sha256": "d69d095bed5b10568c7a8a7e14cf1550db6f394eadc0097d2264ee27567f2aa1",
    "task_id": "d26d24e9df8d6e90f06e0379fd2b15973c134528b7362dfe4112ae2a16352728"
   },
   {
    "admission_sha256": "f60357073e6511dde7e4d2df1cf7c37441f731ce1809ff4c2ece6330b41f017e",
    "task_id": "f8f99219a18a371df0b5ece9ab3f71e6ca34801ddf02b1cb24d072056ccb031d"
   },
   {
    "admission_sha256": "282b3996e53679db88bdb8a51e43692b0a3274b33208020bf140e708397a644e",
    "task_id": "6f2ebfdc856b1bde94ef369c23286b3bdb3838c52ffdd3821ebe66e4de746306"
   },
   {
    "admission_sha256": "a46021bb5db33151f14ca7e114284e0a3da906214d99260f2a83b8bb13e8e2d2",
    "task_id": "dbd0adfa86e581b5b2db01ccfe1159be08509f589828c7dff49899506cd98363"
   },
   {
    "admission_sha256": "dfbd929535e10a54df8134dc8eac2e5f9cc99f51c4cdb4434a1cc6641ad63e4c",
    "task_id": "2ba377a78a35d87e5e2274ee864038f5084b1186424c901080b7ab99a24edfc4"
   },
   {
    "admission_sha256": "03324f3aba73c04352ead2c4c1c6d45b78145a33dab47f1266f5dfcd05b8ac00",
    "task_id": "ef1de3af0a039a53978b7ededad12fa3389d99693c4753078bbf5967bbc90517"
   },
   {
    "admission_sha256": "4d7c49401ff8521a5fa517d8183bbea492a5ccf974717dd7cac59a05fca54a72",
    "task_id": "10f0ab3e4a9e8a16cd6fc962cba3bb2ca546cf9896f85fcc45fa8787c2dd1aac"
   },
   {
    "admission_sha256": "5346224d9ee84c539690c5ef2d19c0d67acb4e5639b1a25f3a955ebf0716da3e",
    "task_id": "643c5a6647d87bdd3238dcfdbacd5b75a75f0f955b3fdef9863b7f86aedeeb7f"
   },
   {
    "admission_sha256": "ac3d0fca99a9e68b0e34e59568c42a066e7b1bd738330cd3fc9a563d048db54d",
    "task_id": "6755f115c622b906a4621a29b77e4da954a99385a070bb934f0340d6de00c26c"
   },
   {
    "admission_sha256": "85f4049a107d7772bd5af3d2809034be6e5cdb45f577efdfaaae1b6b3f7b2696",
    "task_id": "650ec37d633c3738648a30bf36196ed245fa82b9fd49d6497577a9b2847721ac"
   },
   {
    "admission_sha256": "cc94fdaad8258c84e3efce1835ac42f52338a0c14d3806930d9861c1a6fbdad2",
    "task_id": "daf78c921424073b354749eed93a64d4a4e96ac3373795e2b520aba38dbbf85c"
   },
   {
    "admission_sha256": "cd469656b21f9cd90cca25f1642c2315a978882a173683ad2cebccf4b4433fc3",
    "task_id": "bb2c845b2a0f61a62b951d3a90aadf3a674d9f8f38f455e9c411853b73fbe279"
   },
   {
    "admission_sha256": "3cbbc002fdc9fcbef0af3094477730ed54866f45d4b483964b6f69cce80affa9",
    "task_id": "c4a597e17c475ad6ff2d28fae2e9813f48164f1ac7ed0b9eecdeececacc541e8"
   },
   {
    "admission_sha256": "58b3f9489222d857179b09489f06b42edc611d719a328fa35fc03482bcd48932",
    "task_id": "6c976b8c52be79bc4335f0f926fce070f73eac35ffa6b076b2ebcf66fa70a4d1"
   },
   {
    "admission_sha256": "8a32d8598d66c7d3c60ba4db40844c786a2a7caccf8f039473521a09ead93cda",
    "task_id": "65a3f5d066f90984a8a2124573a04a364cd3d23c6a9dd170d78adb2ae831872a"
   },
   {
    "admission_sha256": "f13de95e78f732eb491eb12d3f49cb0746febbdfe43e7b346c8b77de95a4be76",
    "task_id": "0626c1b6225cdd3cd14c636d04112c3d66be2201b3d32041143168e0e29dbd5d"
   },
   {
    "admission_sha256": "89c8a8b9d04bc6587e8da98c3122a9e359b3a7d253db971bfb5c5f74a763035a",
    "task_id": "17f7ae2d3962515b5a7922304db41d0bc1f49af3b3ae7941990b4e5e3cd6dbb8"
   },
   {
    "admission_sha256": "a597bd137acaede345419c26d79639c3943bd41010dae24a1eff690f5c6d14ab",
    "task_id": "51d5377c3168c944aee9eb97ab9c2ae843d2d6e4bdd3e84b7fc92f6f41bf651b"
   },
   {
    "admission_sha256": "55218f372c5c1ef5bac8422e30d9bdf61d2d50ccf5342e3136c84c91497a765d",
    "task_id": "2165e07714d58bac16c0171765f19f2a8323618593e1ee8cbfb17bea12a1eee0"
   },
   {
    "admission_sha256": "477fce3d60fb5fbb3a87dd8b28f98ef6b85bd7707ed81fb60a826c208d3effba",
    "task_id": "9f2ad5f00c6a266097e581c73f4d5aca59441c68fe1113ca2ec391fcedf68be1"
   },
   {
    "admission_sha256": "4f6505a06e8824e3ef20dadd730552b1d57708f469177a32098cb9a9a23538c5",
    "task_id": "6cdff1d90dfd539087d2776e4a33406173e899e38b3ccde649493caa9369430a"
   },
   {
    "admission_sha256": "21ed1c6f8eba87255fc2d132434c7ffa44dd3594cf4d2eab077ccc171654eb96",
    "task_id": "9e7e2bda5b28220099edc9d76b912f69346dbbfe8dbae4255f1f83463b5e848e"
   },
   {
    "admission_sha256": "1129f2290e2e41752ce0323a3bd6768ba72c0304e70f4cf18a8727e5219ad325",
    "task_id": "557e90483802b5eabc55d70b4f5e0852f2599b3b11f32abced72762ceb62f59c"
   },
   {
    "admission_sha256": "59c400f4626d561c716a47eb44736bbe7cecf6514748b0fb25eb0bf36a5daea4",
    "task_id": "d440306a56134e5dc6f4ef006714473385f99393b134b1e2fdeedb8fdaad808d"
   },
   {
    "admission_sha256": "0006c1f465edd26e285b4c04a7b47154f98c3bb08e33310fc282f46ffdf231cd",
    "task_id": "1584165c54c7d8497377725e08bc49384270ca4dde1aed095f4f251ecf0295d0"
   },
   {
    "admission_sha256": "a33371a2421293383d69a6a136caf04d1c2651ac316d6a7fe16d20c2eff9de56",
    "task_id": "d2cb473c5c44a00a6811bb6c365582a0a647e9616c54389d0e60bf79b45b439e"
   },
   {
    "admission_sha256": "38f729352fcbb3e0e8f4f5974f61e9a923d68f2a2d530d3409fd06de41912f60",
    "task_id": "a88e80fe3a4c4aa396c849b2aaa0d91fc88b4c45fa449361c2ba5d6f0b5ef141"
   },
   {
    "admission_sha256": "5179b79850756da04903a80bf15a643b622066486fae5159ee009a93be943d15",
    "task_id": "2ae8c1c65bc0daeee9421e32938642d8088bb019207d770f2646d4f06ffe6ed8"
   },
   {
    "admission_sha256": "98a7048a5c631c8156ef9b5326b034aee19d08aab5763ac0ef5f6f84149a79a4",
    "task_id": "f94a348185e0992e971ad899259e236fa70c174e04ddec4f503e19362ff26b56"
   },
   {
    "admission_sha256": "43b7e6db18b2e693905f0d020bfb7ffff68bf12ff4f0e682a672294396b5fcf7",
    "task_id": "dbebf92f83c2e084cf1c9cd1ef335cbb6f8f27bfb4d814909a1f93930f56191c"
   },
   {
    "admission_sha256": "c66ded5a9d1ec643c0e227ffaa5c15c52172866e33b5ec484726e0dc88796999",
    "task_id": "e9251c941453d224e126f90ac6a5ddff61c8990d807d86286efd8aaa98fafbcf"
   },
   {
    "admission_sha256": "d9c5f48adc25f8a92f785625310088ef7603ceab321de00d63131b51a1c702de",
    "task_id": "3a62740a8d167ee4152caa59cd76836641024499490787bd958fcc559cc9ade4"
   },
   {
    "admission_sha256": "82afaceec3b7cea70115addd8d74d50dc0b5d094236c93d776ee8091abf63522",
    "task_id": "d376afc499a717895ee9efa5b996df5447627fcf7b1f3bb7d591122f35831dbd"
   },
   {
    "admission_sha256": "d15b4028d6744bd7ad7cb7696caf5c5084837ea969d41d317fd0074085ac7433",
    "task_id": "dd4e3ba539a66ab4e9ff205c957345e2abb6ea7579e070a98c8d83790c488938"
   },
   {
    "admission_sha256": "bf50fdb67002c17ee30ae6683f7366622d139d972d603e9ba30e9a661009eda0",
    "task_id": "6f173dc5463b844a217a788360fb042b249bbf537bd15a5aae5af6364de8a323"
   },
   {
    "admission_sha256": "cc25d530fb93b42243e51869d24c42807c59465f14ce9f8e0e95888e2c11f7a5",
    "task_id": "17af830b7e312c12fc7dba96357ed7f566373f1dfaf64a8fe58e7233c9328c0f"
   },
   {
    "admission_sha256": "240425d3b39be9d2462aa2d8f796f51f3a41d8020948feffa93dccbccceface4",
    "task_id": "1467450fa543c5b3d229b9799e56029649bf63ea18fb8282c3d1e5af7660befd"
   },
   {
    "admission_sha256": "456ddc1f6ff3fee23174037db54ef184caf731802104f05d01c30e5e886e3218",
    "task_id": "f17fb07b7134ac7ba12e82e3b1bfc2186dae25a22865125c95e2f5425d0c65bf"
   },
   {
    "admission_sha256": "9c2633293500458e2b730a875a0ac7507195d870494b0bea6d6430256dbe5da0",
    "task_id": "c1c57e5bfa04c5711bc183176883f8b756a7126f02f453e9e87e4a9ffa3750dd"
   },
   {
    "admission_sha256": "d830158fa0950fa75bc615912fcea27efe13c227e2a20db3c7d309faab742201",
    "task_id": "254cb3d02e78185943668f2b6a286861b0331ad248ef70570b55c35ad3199e73"
   },
   {
    "admission_sha256": "7de7828a4d86bc8d2d095a0af3ee9185d87168cb119058267351e2b5106596ae",
    "task_id": "b615bf2c9ad5c11d864e9b224a9900053583616f9198d57def21fd14f50fa123"
   },
   {
    "admission_sha256": "7c08e6a9976e1f6c8c91f91b694b1dd2a6a40b233bf4542199194dd826d69ad1",
    "task_id": "7f0283b76218af26e3814c16e751367935323585bf7803fb0a5cc55d2d9f1694"
   },
   {
    "admission_sha256": "a012aff552f3745fb739c882620c5177500193a82b6b4b6bba75ddae693a66b3",
    "task_id": "bcaf170472342f98cbcc8cbb409d488d4a29a0e9e56244150a9c5e32dcdd85a7"
   },
   {
    "admission_sha256": "1bcfda598caa33fde55003792e61dcff00f3faeca645eb47037908512b1f065f",
    "task_id": "4b28ca539a18f695e188274c76b9d34ac2ef1506c6a9588ab6df30bb9597dc0d"
   },
   {
    "admission_sha256": "2ac778307ff8be909b854d6a2ec348c6b0e6e4071c218be9bb908ced93034159",
    "task_id": "77eb21c9e5f6c96e6e84fd195287e45ae74abd5f71c1be4d2b6f08df06c61c47"
   },
   {
    "admission_sha256": "906dfe9f21e2dccd2d52ee33429f146e356598e7f56ce37836135c54638ccfdb",
    "task_id": "8152065ad27babc7aaa1d6e3e7806016050ca4cb961b56f30b522cd1f9cbc9bb"
   },
   {
    "admission_sha256": "e8e32843d4aa273f2b4df5b83212fd51ff099724852b382df80d174b4aa6a3d2",
    "task_id": "ed2996d75339c125ce7d2e361eb56c55d934115281b960cd983bbc73e473077b"
   },
   {
    "admission_sha256": "97eb2611ec1a5a67a04ecdb4f8a6ca72fbcec5dda83f44bd02437e6e8e23d88f",
    "task_id": "61c4299e2db414a4733510bb2833ca697fa3401ee9c7c5d37a8f112aa1306877"
   },
   {
    "admission_sha256": "00921e7e1b2a69702ed5f4ebd3e4f3f1974af8d37c60a02a27057df4498ce891",
    "task_id": "bbcfdc7dd0700462c9261d95fb6eb1b13e24d2e1f6ecbdccaaeb171d2fc8e0c7"
   },
   {
    "admission_sha256": "4aa9c68d123846cf4e3d01787957f16e68791e8ca6a5526659b13ad60ca87cf6",
    "task_id": "c9329623ab982826ad353e80be97e4a03e35b46598fac7585206776cab499075"
   },
   {
    "admission_sha256": "178dd7ebb5a6194d058fa772ceeaca2eae0c0d4ff3968c3cdd82bd5245108e27",
    "task_id": "eb6eafd5633cd628d12d8a2fde6276e1c00eeb03342be8990848195ea11615e1"
   },
   {
    "admission_sha256": "6a526b6211e1f9f2554089e716c8e928857624df52eeba6d11c297b631737ac7",
    "task_id": "842eab330f748d2be004df1eb28563440d937d2d7f33de433d8e4f52707e5613"
   },
   {
    "admission_sha256": "741e40808469bf25c451805863f2cd94317a31868079e963e986471eed91ccf2",
    "task_id": "17924edb1035bba8c5425b6dae8359ea7157a96a5bbcb123f15cf935cf4cd491"
   },
   {
    "admission_sha256": "540c4f428666e55c0d5c3d670067a5baa909b0ce4c1649f888744d70a46d22df",
    "task_id": "0c056374b2f2d31fb850014daae4fe8de6d942f1dcb28507e0ca53eb548b9f45"
   },
   {
    "admission_sha256": "2fab73b78c74c8e80a3e5a16db89d78c3233702b900887b8059054a289a2cd57",
    "task_id": "78e518acdc813f93ef362e9b798475b237ef11caf9cb40176b9978858d4da2ec"
   },
   {
    "admission_sha256": "7e447f3a20c1d6ddafdf9669888f39ff2a87307e525dd6d46489743baf0d52f2",
    "task_id": "ca11bca64cc309a78f5c10654c2415ee8832ad65d738dbd452f28177b4c2414c"
   },
   {
    "admission_sha256": "65b618e611183a7ae06a2de3633eec0c9abfcebc4b83a708c2d568c11e82bbf5",
    "task_id": "e6f9a85d583173434203c4a4735a2ed0de713a672c5e9bcc8b9ae511fcb5e154"
   },
   {
    "admission_sha256": "a4be1b483b79af626bee7c0883c5eb00430e502ba66ed3699e04c948376f4310",
    "task_id": "f564a69990a2364e138c9bdb88ad9fcb32e59a1822b98a25d80a5fcd14d08e13"
   },
   {
    "admission_sha256": "3ebe73c91b290991111df8b905abc505da8adadb02e3d6bce8947061b7571db1",
    "task_id": "9ec8ed3affc97798b55692dfb71270bae78ae923aebf0d63a4525f776b9e6b15"
   },
   {
    "admission_sha256": "ad521a4448b7e47a65237fecc651dd1e03c9970f33f48e992fcca048250ffe40",
    "task_id": "92fbed84de148788ddf911817eb6daf25679b1814af1a79ba924bef0deacedfd"
   },
   {
    "admission_sha256": "1917d2e20e5932f4dd8bd786ac99cff734556f15e0a888ab30332dd2947eb0ab",
    "task_id": "a1fe94c2a58ebc5b6e8913c5e2e021e9b27aeec5b7c8fc4756987784e9625142"
   },
   {
    "admission_sha256": "0ec8e7f7aade1c782087b16a70aa487f791d81e6ba8bfac57367c6bd992193ce",
    "task_id": "62035b4866e962208abea8abef7eaf47a2dc11829923aadc903538c9834fd857"
   },
   {
    "admission_sha256": "16cde55391c71ddfa87fcd06506aef819a42481fd5a7127e3f7bd341a48641ad",
    "task_id": "0b714011b89036aa45ceafedcf9dddf81bc237a621a33f0b088bd3305c4553ee"
   },
   {
    "admission_sha256": "1224ff3a86d8d8b011b812468cbae50185bcc536889a4f842063c36a8378643b",
    "task_id": "27ceb4df0f1e892f6aea9e90417536d22762bbbf38813cf325e1d9af47225bb6"
   },
   {
    "admission_sha256": "8e42f8ad0388b58c4de7b1d4855175edec7d474ebbf7a4a1d54141f79595aaed",
    "task_id": "01ef4e41cf91f0d09315f14b25f66dc96ff53f7a17d4b98a6cf98ba7dc646ae8"
   },
   {
    "admission_sha256": "c9a587e7c5805ca5469d07d21e77d5b8a48952835b28fa5e5444c9c03c248bea",
    "task_id": "0a1c14c325d1dcbfcdf01ca0eea413440ffe05d49df75e136e5eb4e964ce3456"
   },
   {
    "admission_sha256": "a5887e186362d5375cec37765e694e80e97dfb30200dbec783baf2e40b512db8",
    "task_id": "e4282e2e7376065b3e275a9273e7f9630a67c69e8f8a87353cc2caa21dc81b27"
   },
   {
    "admission_sha256": "1e53300af3e55af8c525e10ca101b0237fba7a718786eb5a6a8069d3bfb02452",
    "task_id": "5d897115e48c8983041eb4cc402286e58296167af49dd030d8ea136e77d60321"
   },
   {
    "admission_sha256": "4f97fecd88f4d6f44d776bf426c7263f6769af1257866f103545f3dcc0b78590",
    "task_id": "03aa57f7a3ea8bfe15aae5b19727ba3d54cfa0e6a25e9c4f0e9388d89d48fa48"
   },
   {
    "admission_sha256": "76ce8c4b89c15eba127643f7076804d2cb6e9f0275ff39dfdf67ec6bba1af417",
    "task_id": "88a6051ddc73bf38ee59ba9bccdd76a05236c6a3a0917324db478f6a10e534a7"
   },
   {
    "admission_sha256": "db1de19fc6571348e69516ae57dc3ab967a3c1243cd438e7609bbf3cd4585016",
    "task_id": "4a9706b1d585b5e93ba72e838145cbdcaa742049af1d399bc8269c85a0693b7c"
   }
  ],
  "config": {
   "objective": "vanilla-dual-clip-ppo-seq-mean-token-sum-norm-derail-masked-k3-kl-reference-overlong-soft-penalty",
   "old_source": "trusted-training-engine-prefill",
   "optimizer": "torch.optim.AdamW-cpu-fp32-masters",
   "rollout_correction_bypass_mode": false,
   "settings": {
    "adam_epsilon": 1e-08,
    "beta1": 0.9,
    "beta2": 0.999,
    "cap_policy": "affine-unresolved-zero-reward-active-v1",
    "clip_ratio": 0.2,
    "derail_mask": true,
    "derail_tail": 64,
    "derail_threshold": -3.0,
    "derail_window": 32,
    "dual_clip": 3.0,
    "entropy_coefficient": 0.0,
    "group_size": 8,
    "kl_coefficient": 0.02,
    "kl_reference_checkpoint": "87de8304e9d33dde39b81a41fdbbf306426ea6fd993b7d85332e6a48e8a4bbae",
    "kl_type": "low_var_kl",
    "learning_rate": 1e-06,
    "loss_agg_mode": "seq-mean-token-sum-norm",
    "loss_scale_factor": 8192,
    "max_gradient_norm": 1.0,
    "max_tasks": 256,
    "normalize_advantage_std": false,
    "old_actor_logprob_tolerance": 1e-05,
    "overlong_max_tokens": 8192,
    "overlong_penalty": 0.5,
    "overlong_soft_start": 6000,
    "ppo_epochs": 1,
    "sampler_correction": "none",
    "sampler_ratio_lower": 0.2,
    "sampler_ratio_upper": 5.0,
    "temperature": 1.0,
    "top_k": 20,
    "top_p": 0.95,
    "weight_decay": 0.01,
    "weight_decay_policy": "mcore-bias-and-1d-excluded-v1"
   },
   "upstream_revision": "a2ad9f6160b03ff2d47e59832bfb6b289f37c917",
   "version": "affine-mimo-reference-grpo-v2"
  },
  "data_sha256": "e85e7ab95a6097da9f49bc5a7beb06adb64cefb47d3e8a84f78a413b825e45a3",
  "derailed_rollouts": 0,
  "epoch_metrics": {
   "aggregation": {
    "loss_agg_mode": "seq-mean-token-mean",
    "loss_scale": 0.5635,
    "mode": "seq-mean-token-mean",
    "signed_config_loss_agg_mode": "seq-mean-token-sum-norm"
   },
   "cap_rate": 0.0,
   "cap_rate_by_level": {
    "unknown": {
     "capped": 0,
     "rate": 0.0,
     "rollouts": 125
    }
   },
   "cap_rate_math_l1_3": null,
   "capped_gradient_share": 0.0,
   "capped_rollouts": 0,
   "gradient_clip_applied": true,
   "logprob_all": {
    "rollouts": 125,
    "token_mean": -0.09328864476643789,
    "tokens": 143484
   },
   "logprob_capped": {
    "rollouts": 0,
    "token_mean": null,
    "tokens": 0
   },
   "logprob_pass": {
    "rollouts": 125,
    "token_mean": -0.09328864476643789,
    "tokens": 143484
   },
   "max_gradient_norm": 1.0,
   "negative_advantage_gradient_share": 0.0,
   "pg_clipfrac_note": "actor-as-OLD with one PPO epoch: ratio == 1 exactly, so PPO clipping never binds",
   "pg_clipfrac_token_mean": 0.0,
   "preclip_gradient_norm": 1.9380725622177124,
   "rollouts": 125,
   "rollouts_math_l1_3": 0,
   "unresolved_gradient_share": 0.0,
   "version": "affine-mimo-epoch-metrics-v8211"
  },
  "group_count": 125,
  "input": {
   "checkpoint": "87fa740811e75e4c221258ce6cf2d8cb1db1519d69aa0a65b42cb5fddea5a499",
   "contract_sha256": "a043ed72cae8b792869526ae6f3760280fdcebcceeb549c688324e8321fb7378",
   "epoch": "affine-mimo9b-grpo-20261013-e000354",
   "optimizer_step": 122,
   "run": "affine-mimo9b-grpo-20261013"
  },
  "input_checkpoint": "87fa740811e75e4c221258ce6cf2d8cb1db1519d69aa0a65b42cb5fddea5a499",
  "input_documents_sha256": [
   "31107b198ff5d985c4c21d0d04465a7bd45ef4a94deeda256119dc1dd5b1c27c",
   "d94b9a451e6d84230afcd2b22101ed35433b127e8ab38416a2a9399c267200ef",
   "e9787cda658f9d465a07d2e1a1301adcf8dc0ee1b12c6a01e9aaae6b1cbc2fc4",
   "2a3c3cb151f38cd8534ffa72e25c89c742e0b76bd60bc2aec8f9d395b1eea79d",
   "a2a4b0b33a2f907b1486979295716c24eee4144869ec6b456c413f4cb8d24dab",
   "5c011342071268aa046e643fbf4e5a360eb951f38b3eb3b916e63dcb2df56c62",
   "11b0f90b3193ec3c6c71dbb65ae2561b58b5a67fc1785b7d45e1960914348136",
   "18a8155f9442982c43e3192cb6b0f0f2e478246509393edd9b7a92d530a46e70",
   "bf4db80fecc6c6eb099e436939ec6051bfb24d13e73c32918ef4c7ca29413450",
   "1a1b32b06a0acbc1d15ebe62af1db061baca59a85df075b6f0669a53101985ee",
   "4a90360cc0e8be1df012706628a9bdf8c02e0a19af81efa78b1bfe1886517e7b",
   "37d593617587e50336cfbe3e4b2fa19442a12e34e4c2284371b2ca8d37c69758",
   "23a1381c5d3d6cdb796430789dee5e90a47f975aca9086687ffb017c0bfd706a",
   "d0f6a37bec04202979e9a00e576f0c15cf2c04697acadee285f827b1fce2cc73",
   "1ea80eaa5380895b8e39d44fa0d0f937b099a34031a487fed34d527aeba8316d",
   "ca5d42833d3c3add7c21af69582a677a589cc50c3b529dae7e00b82d517ae7b7",
   "02fb137ef4a48bb8eed1c70857d72278e1703295bfa190c5c26d75323055c506",
   "a3135b7bc86b6a23eb3679b4bac3fbc570c05e028fffdae777d3ff1ae0f45325",
   "eb6f39211cb007a14e4e442ba99eaadd9fe5574d90299a45b4d6f61c5a79a755",
   "bcf0f6ca7975850f89e1dc3328a46aede722e0df1229b18c00bcd31b8a49049a",
   "7ebeee25baca1c2f8f526c2502344bd5a233fa2311ad7db3d5dd644136f4ae2e",
   "78038e79bfa1af0f141ecf050cc8537ddfe0c978cb0e9493690e00f3a9101d43",
   "04084384123448e14afc5643bb9047e0521f5fbe5d43c52b563ddf1b64d19b54",
   "e9f2c86d43f04bfe7b8888ac6d65032885b8286f795499f265b3e98d7861d7d6",
   "f8e4925144460f8efd9c60a39bbed9154626247fb2606b9c9aab37595cea2ec6",
   "d74aaaa9e815838fae059b7d384a3a43e37d76cbe883fb8c21e91d44805d7d2a",
   "a049eeb902d6185cc4e1de83d91efaf14fea5cb30605e447df6c81dbc4fc276e",
   "cf55f58301c1d5441105464665e43b64c7afeb419ebb1ca46c7d82f5ac630252",
   "c086c4b1cb6a1ab7a8e5862b1bcc10a9a34b6dd974603aed434035a7bf64d651",
   "418fb132346c76ce5d02042b51794e089115760802c5b5594a53a02970e642a5",
   "7fad1afe47c377be5755e99e183eac2fe7f68eada863b198b00eb57c24fe065c",
   "42f7c127f4602be4905bb6ca1715bc444bc30d653674d0c464b22f5513a7257a",
   "36496ab8cfa1381dd51f80d39d4b959cdd87ddd5b5890dafc8e15b8ef4e70525",
   "b742a3a9ddeaa04c975df8547d3bd9a740a45f4348557352442e9b3f150f2db2",
   "1b8e754fe83ff336288da4749b13ae537e0daf1237d05a408e9df9b340e7f6ae",
   "250bf5389f019d761ce81250fa7401fe3de36383ad191c744d00cd4335eadf90",
   "da47736b2b68ea3ee43fa9dd2672b36a24d12337947dbf89a4f038e7787daa95",
   "c81a9ee75ca1fe9b0ce994382edfbd84e9af7b173118d91378f102ac380e98aa",
   "6bcf60281fdcefb491a1b75c032a13aa29dc82f991bac0023091b89c84c51c08",
   "3a176f4753a65e5661923e07c5b28034ffe2b240b18404a76b998b48c3054caa",
   "6628d25808ceaa146c8ad3dc212e79ceaa7d358bea525574f5e3c06fc2a71976",
   "78f5156d188f60043f13cea95776fef0ff6d7e2cf7992bacd46b32b0fbcf5854",
   "c58533388c1de8ad523f0e1f86938bb758e10efc56c9f8edb40f7ef15f8f5900",
   "6c366cb57554196eeb09ac5fd009ae355beed3318196192404b6942943af42b5",
   "9b8400d1bc49def1f29bfd3adac3ec0c4200bc5fe35b8a5f69ba255d462ea292",
   "a90d24c9fe9164faf9929e08e7c81698933088c4003914d1015a413f664c1eb1",
   "232da2b35d56b7302cfd973b201ed18e8657fa531f6794e460af8ef67fd99f3c",
   "a6114b8e54c336f9d91cd808261f1eff22c1b06b68587000c201ac8d76606265",
   "e6d8201eb3f87f588e062ab3de5754a8c2f9e9e4ca9bfa5589c03127a5fb05fc",
   "94606cdda1b435f193945420470117de33352b7cc1c557288d47ddaddcc0fff2",
   "6c62441ae749024a85292804af206548fd2030e363dc54238b06e49fd23b65cb",
   "525500db6798d41f3d81482ebdb96182dd43fba98b4ee9149b7c4fac0e85643d",
   "38a6292bccea890c863e6de34a96f7fb8baaf2f005032b60d56c2ff3676d9f60",
   "b55fdf6d5ac4ae31c30f3f7a500f0ba3dcff56df6e8bcc377af71a75c5b3a8cc",
   "2c8d85ae6f674912b75c7c9b51014159e4c497187a05b5f1348057b2e4dc1044",
   "0e201f3a90f46224650026165d7d79fb4351aa3beb67f22aa625d2d59920dd42",
   "a2166e4eff1f150463ff23d55ab1289bbae29f75cf868e05ace36e2b67cb7c3f",
   "ece3ad7a6bf8081f4bde4ace461e93be2906505bbd4f196d4b3adf760b4a1e48",
   "c977e29ed1b1a86f408467026b62a33672fc189be313c8e0455cc10e199dcd3a",
   "85e59c5fdd4a437bfd9ad7cca1b196932243d8a91cba9131fd66098fd5bc7e01",
   "a2037e3843b1010868d18da835438b40a38a3958a14b1a36152f4c4c80a288d7",
   "8dc0a183cabce10cc0b92e3cfb0e9bef0e83ba509ff06e1b4ae8076b74c9bf80",
   "24dba7f951e7710673f15e53400d28f11b7d54add417bfd61563862f423db2c7",
   "de8ef5793088cf078564ef763838f57cd37c413b0360c640baab809c1b9518e8",
   "269aee86db4a40643795e2e634aae4dcd5bd2b72fdd538ba39259cb8e48435c5",
   "4683c196cabb50020790f67133da4e9b5e65ea7d96b6aa90f6175a6508f6fcbf",
   "026a9cedd9593044122e9b52740120065b400d0de94517676631ae2506e140e4",
   "d98ee2ab45c08876e1476da14553e994bd9f4a64099926186e530f7f41df70df",
   "c098e55a460661f9064c9aa92ecab237f2a754ae55d1b5103ae0483fe7f94272",
   "58372d99cbf75d0921c481c213c4b052d90cd59d68498112bf208a2d8f1f0b40",
   "3b4d621778471cff9013d8d33fdd42fa9bb426a288483de7877d2d014666363a",
   "1adc12f55a9ee166faf04faa6835c7bfe13c5d677e5773753eed1f90fec05fc8",
   "ec757abc4ccd9b7260d53cc67eebcf245f2f545dc5ba88e46bce2794fdee29f7",
   "1e0885def6c7c62f88fa695b89878186817fc797fe4b05a46f63458fc17150e6",
   "49f9bcb38f5fef30f090839743fb02091da32a4701cb87454992f1f4f6362a65",
   "4e32437e3c0f09498ee936bbc26f9306e3d107340f29475921c132c509c93222",
   "66294db4cb08824e258f2c33cfc5dd8da04a7f52c60cd79cc57276b111dfc25a",
   "a2e50f2edea5224c3d579ce1612a45cd659c48d44a5cc36e019cada0d3896763",
   "d8bc3220c5be9166b8a521f926fa57fc8aa3a16beb5dd7d1cd735401bf712895",
   "381252f7c3a1ee00da009cebfb0dcfc119db457c17ada31a18ce489a42bbbac1",
   "13e94f97ac6dcffec77b2f8c71cab9b22f4217b849438590965df04c2268bd39",
   "ea35fc2c225897ef0891f7ed7c88c84a23552dcc1516e94cda49dd977c580661",
   "7eb20053fab0b4e7e2b9cef351a248611aebe3ffb09a8dae90cb8c3f3f8528bb",
   "943c1457931cef3728b78a69c667b50bc307c8d8cbc3507cdf5f7e2276804e64",
   "fd80ad734ec5b9068c089da7156943b423d3227a388d1c36be690e75244fc671",
   "827b0a112925c654ff682ebbfef82fab49d9356fbd59c2238dda4f79dfb9a9b0",
   "4a7f157994c28bb07f99d36d597928bf4ff32447c9ccfdf88cb6e5fe94cae865",
   "954ca9f325a497972bc309158284e198a4a1c078be5a7ad2acbb70de2cd43bb2",
   "4cea512a35258ab167cbffce94da20b27918a794daea3bd6a15ad9e6976a39f2",
   "f037dfd4f23e349088a6c177a52f19da0b42e0fd0dc73c05194067c87b65ee8c",
   "75402701b5947bb1281c0c4a0e0847c660cdad8bce0e68ad62981fb4bea7f3d4",
   "6f91c4cb3f5c78115030548ba974b93a1993cc7d596248a5fe2bb4ebc85e979c",
   "369708c9bbac8eab3063e1fcf02a7048a8ddeda56aa4a6711e3482564de43df8",
   "65be87a00d9cefa88431e7f092c68f01fdcb4ec1e710bc59c97ffac5ef41cf4b",
   "282ae559e8374208a6852cee5e32fb0476cc61c45e988464cfc8dbb19f55d8f7",
   "cf2d93c8a8e7c8cb31fe7ff31f7cf790236e0a81fc0150a4f4e7f6710c895752",
   "891bd1d9f4892acd5ce2cfd62efdae7ad9204e39d3b6ae30ba48f27ba178caee",
   "338b86003ce276d6966a26c6067c35f6b9fe4ccbfd1ce8f191191326dbba6647",
   "480d516f43c2d1913e98b6e59af5825741c06e22c4c4ce31cbd769d5e93e9086",
   "9e782556aaec5b882a280cc8a0061ffabb9f8e6fe49c848e06302aa2da0af90f",
   "108ac44311c006ad76a7d8e57eb62ac1737ae11772c1e863f0b43ed661f490e3",
   "4c44ca25e2798260909c29e6e47d29b3f6b640bdf684ff592d6e893aaefb0e3a",
   "84f823800d7ef292f211d31e317accf9daa895331f5f8866753dbd803b9de4ff",
   "c8d13e08c90d702ccfe084a3d402475619e2e54747dbda986afa32d6c55d32ee",
   "de3312cc05cb56ed8f7bffa3c8cea14467a60fffc2efe62546a0489c3c610461",
   "b97555efb020cd238413e76426f8a882f70a04e8b0d671c927d91702090af74d",
   "652ec3da588dd92786098680354fe6ee9c5e9ad377a6d9fe7244b85ba7db6b1a",
   "8c869d3ba8fca2e9e29974baaa9a4994e2e8a91fafdabca0e8c1c3ef4e4d56aa",
   "a9d103fa198470436233a32a093a619afd7cdaa2820514afab12f3c786e3b1de",
   "e68312548a2b366d831721705943597075c5adedd97e5b68b6140d16956f18b2",
   "31f1031a66381d21fa32fe644dea92d8f71317bd643f030fcee25c45120a8007",
   "6bd74841f7736287d5184f355ee23227c9c7590fcb6d4314bcaa9022f20c51df",
   "2a6d731e5c4793091076dfdfd8675103e1a87acb51c1d2abc3bbfb15e6e9fca2",
   "4391b453fe6c3274163e52d900579cc0bee1a43db0864cec92adba09197d7e18",
   "02582fe6ef902a931dc0edbbd152c37f0a0a386525177ad21ef3b7f3c2ae8b6e",
   "aae62ded98b841f60a688ef2173c22f7df30191d18ab3418b910a48f275d4140",
   "781eb600191f8fea918ca099fb4568a853792e87d5913514b7857bd4d19c17ac",
   "395011e520fe804de04120c559a770673b7b7a4f273019835c001163a4228ae3",
   "a0d3e495dc2d75aeb962550868b7aea75c99c62c1bf76f1553319fa8b7e0ec47",
   "d429bdda7d943453c8fb118426c27870c99747cccf0cbd68cadad0367edb48f8",
   "9b73c78fd8b4c03f594a4eb0b7a82c605e75c147eb01016d4a28b627583a8fb2",
   "89d199c320f5d20e0ba327772f32879b1eec693a8b32daa1fcebae773124806f",
   "5cfd33e199666164be3f6242d0b5118bad44be62042d7cb0b2d35f4c32a29a36",
   "fdb917a686d65906c89b52ade2d70c89abdb0d78d37cbcff476b208da4e18ef6",
   "f4664905fa1c0e7bed08f3728838e3e57accd84f729dfd6b2b70e6490d3132fb"
  ],
  "loss_aggregation": {
   "loss_agg_mode": "seq-mean-token-mean",
   "loss_scale": 0.5635,
   "mode": "seq-mean-token-mean",
   "signed_config_loss_agg_mode": "seq-mean-token-sum-norm"
  },
  "loss_tokens": 143484,
  "masked_rollouts": 0,
  "maximum_old_actor_logprob_drift": 0.0,
  "microbatch": {
   "count": 5,
   "engine": "packed-dp-v83",
   "max_rows": 256,
   "rank_tokens": [
    32768,
    32762,
    32752,
    32652,
    27150,
    0,
    0,
    0
   ],
   "token_budget": 32768,
   "world": 8
  },
  "objective_loss": -0.5632560248486698,
  "optimizer": {
   "betas": [
    0.9,
    0.999
   ],
   "eps": 1e-08,
   "foreach": false,
   "fused": false,
   "gradient_accumulation_dtype": "float32",
   "implementation": "fresh-adamw-norm-gradient-direction-fp32-head-master-v833i",
   "lr": 5e-07,
   "master_dtype": "float32-head-master",
   "momentum": 0.0,
   "optimizer_state": "none (fresh zero moments every step)",
   "optimizer_step": 123,
   "projected_changed_elements": 34310311,
   "sgd": {
    "adam_eps": 1e-08,
    "cap_fired": false,
    "cap_scale": 1.0,
    "fp32_master": {
     "fingerprint_seconds": 16.262460152851418,
     "finish_seconds": 15.86553656985052,
     "input_bf16_sha256": "59f6b2a608153fd01de8eb02f2c09e1c0274499d147db5f5740adfcfa75b1fc0",
     "input_master_digest": "e16e430bd6761a2d245210445f32714efae4ea715a653111de80324a287a2495",
     "input_store_entry": "59f6b2a608153fd01de8eb02f2c09e1c0274499d147db5f5740adfcfa75b1fc0",
     "master_bytes": 35815213056,
     "master_tensors": 427,
     "output_bf16_sha256": "8fef0abd37d5072b67254e37db613cfae12a9fd80c1ab6c16d9fb5b835fd6f46",
     "output_master_digest": "4fa713955528282a3d69b44eb9042a11882bd3d1f30a99242cc06e3c09c54d16",
     "output_store_entry": "8fef0abd37d5072b67254e37db613cfae12a9fd80c1ab6c16d9fb5b835fd6f46",
     "prepare_seconds": 24.54028800688684,
     "source": "disk",
     "store": "/workspace/affine-mimo-trainer-prod/fp32-master",
     "update_norm_master": 0.022355193202520586,
     "update_rel_master": 1.716421062594337e-05,
     "version": "v833f-fp32-head-master-v1"
    },
    "frozen_tensors": 0,
    "optimizer_step": 123,
    "projected_changed_elements": 34310311,
    "raw_update_norm": 0.022355193206372338,
    "seconds": 52.19926381390542,
    "sgd_direction": "normgrad",
    "sgd_freeze": [],
    "sgd_lr": 5e-07,
    "sgd_mode": "fresh-adam",
    "sgd_tensor_rel_cap": 2e-05,
    "sgd_update_norm_cap": 1.0,
    "tensors_rel_capped": 242,
    "update_norm_bf16": 0.013469831300909809,
    "update_rel_bf16": 1.0342072173130343e-05,
    "version": "v833i-normgrad-fp32-master-v1",
    "weight_decay": 0.01,
    "weight_norm_live": 1302.4306034051544,
    "zero_gradient_tensors": 0
   },
   "state_dtype": null,
   "weight_decay": 0.01,
   "weight_decay_policy": "mcore-bias-and-1d-excluded-v1"
  },
  "optimizer_step": 123,
  "optimizer_step_before": 122,
  "outcome_counts": {
   "correct": 125,
   "unresolved": 0,
   "wrong": 0
  },
  "output_checkpoint": "ce3971dc21a17a964c03273b68a969984dd43503094937e56bee8cf1280b1853",
  "preclip_gradient_norm": 1.9380725622177124,
  "rollout_count": 1000,
  "runtime_context": {
   "manifest_sha256": "a9c6cc0406d0b0186bbdc0283a6ab4759191ac5046e94bac177d984d61239855",
   "source_closure_sha256": "f792352b040721a5ab86c57eb13413876a426074190d5a2385c89e876e917875"
  },
  "sampled_tokens": 143484,
  "status": "updated",
  "support_renorm": {
   "inside_logq_mean": -0.08514300056194546,
   "loss_tokens": 143484,
   "max_rollout_outside_fraction": 0.012626262626262626,
   "outside_fraction": 0.0008223913467703716,
   "outside_tokens": 118,
   "rollouts": 125,
   "rollouts_outside_over_0p5pct": 7,
   "rollouts_with_outside": 60,
   "rule": "sampled tokens outside the replayed top_k -> top_p support are dropped from the PG token mean only (still in KL and derail); files are never rejected for it",
   "settings": {
    "temperature": 1.0,
    "top_k": 20,
    "top_p": 0.95
   },
   "version": "v833i-support-renorm-v1"
  },
  "trained_rollouts": 125,
  "trained_tokens": 143484,
  "trajectories": [
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.16024644672870636,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.03362162411212921,
    "loss": -0.004504968877881765,
    "loss_tokens": 249,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.16024644672870636,
    "outcome": "correct",
    "prompt_weight": 3.21285140562249e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.14027118682861328,
    "support_outside_tokens": 1,
    "task_id": "47623a99cd3c84cc3f5b106c5117e6a0d4de2957aa7628ac468a7777bc3c8e9b",
    "tokens": 249
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.061199888586997986,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.011028203181922436,
    "loss": -0.004507005680352449,
    "loss_tokens": 840,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.061199888586997986,
    "outcome": "correct",
    "prompt_weight": 9.523809523809523e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.05268268659710884,
    "support_outside_tokens": 1,
    "task_id": "febe060739338ff7b01afd1e01f1c31072fa1095f3b379a68b7ecbd6807790c6",
    "tokens": 840
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.07703123986721039,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.008594038896262646,
    "loss": -0.004507225472480059,
    "loss_tokens": 1276,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.07703123986721039,
    "outcome": "correct",
    "prompt_weight": 6.269592476489028e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.07347230613231659,
    "support_outside_tokens": 0,
    "task_id": "6e13d4931d8cb01e03dbdf22d0952a61aca999fa63c232632a32a50cccf2f47c",
    "tokens": 1276
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.11172454059123993,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.013570164330303669,
    "loss": -0.004506776574999094,
    "loss_tokens": 580,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.11172454059123993,
    "outcome": "correct",
    "prompt_weight": 1.3793103448275862e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.10688918828964233,
    "support_outside_tokens": 0,
    "task_id": "0b33db1a5b7ab3ea5eaf737d908c78452d416b5682203a8ae92dd2b9aeb02630",
    "tokens": 580
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.10130015760660172,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.00808558240532875,
    "loss": -0.004507271107286215,
    "loss_tokens": 1673,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.10130015760660172,
    "outcome": "correct",
    "prompt_weight": 4.781829049611476e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.09200634062290192,
    "support_outside_tokens": 2,
    "task_id": "da1e6903b4411949cdc2e27f2596bc1eb0a51e2477da503f3e243b2f9ad517ce",
    "tokens": 1673
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.05480368435382843,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.01224223431199789,
    "loss": -0.004506896249949932,
    "loss_tokens": 641,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.05480368435382843,
    "outcome": "correct",
    "prompt_weight": 1.2480499219968799e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.05119217187166214,
    "support_outside_tokens": 0,
    "task_id": "5cc075ace2ca1500a2d3ac6f380af314007146cc7f7c083e5d8094ccc675c1a1",
    "tokens": 641
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.1814815104007721,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.0212797112762928,
    "loss": -0.0045060813426971436,
    "loss_tokens": 621,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.1814815104007721,
    "outcome": "correct",
    "prompt_weight": 1.288244766505636e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.15398815274238586,
    "support_outside_tokens": 2,
    "task_id": "728572665e3b46f5bbc74a0bdae6725ff509683850a32d041863c715f8fb0c60",
    "tokens": 621
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.13336241245269775,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.032352134585380554,
    "loss": -0.004505083430558443,
    "loss_tokens": 340,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.13336241245269775,
    "outcome": "correct",
    "prompt_weight": 2.3529411764705884e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.1272597461938858,
    "support_outside_tokens": 0,
    "task_id": "588fa076f9a7e0e5062c2d7fee98c01524c7798edea325f3f0f52956eaa9a35e",
    "tokens": 340
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.09008049964904785,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.007978915236890316,
    "loss": -0.004507280420511961,
    "loss_tokens": 745,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.09008049964904785,
    "outcome": "correct",
    "prompt_weight": 1.0738255033557047e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.08504386991262436,
    "support_outside_tokens": 0,
    "task_id": "dc864ee0a497b0af693911a96d8e566d425de2aaa2bc3538522262123453d1a0",
    "tokens": 745
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.09228365123271942,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.020480934530496597,
    "loss": -0.004506153520196676,
    "loss_tokens": 711,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.09228365123271942,
    "outcome": "correct",
    "prompt_weight": 1.1251758087201125e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.08072976768016815,
    "support_outside_tokens": 1,
    "task_id": "bdcec23f6afa3d95df26b3f33e880fe9ddd11713c8955e53ecf137445de108ea",
    "tokens": 711
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.17455051839351654,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.022698771208524704,
    "loss": -0.004505953285843134,
    "loss_tokens": 888,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.17455051839351654,
    "outcome": "correct",
    "prompt_weight": 9.00900900900901e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.15124428272247314,
    "support_outside_tokens": 3,
    "task_id": "4e53d8ae085f72266bfdf020dcd1bf2b408ac3972e2e174b04f341e4f64f66e0",
    "tokens": 888
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.16019606590270996,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.0080303605645895,
    "loss": -0.004507275763899088,
    "loss_tokens": 3069,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.16019606590270996,
    "outcome": "correct",
    "prompt_weight": 2.606712284131639e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.14767833054065704,
    "support_outside_tokens": 3,
    "task_id": "acf03f84995c27e667b9d698105ddaf4a184ecc0c6cdcba2c48b271c99164c8f",
    "tokens": 3069
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.09402608126401901,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.05951622501015663,
    "loss": -0.004502634052187204,
    "loss_tokens": 250,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.09402608126401901,
    "outcome": "correct",
    "prompt_weight": 3.2e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.059687238186597824,
    "support_outside_tokens": 1,
    "task_id": "c0b7ffc2401635b3c61b1b3d7099332f5d511bd5f46ea13920b88e39d678bc46",
    "tokens": 250
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.1375408172607422,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.012478780001401901,
    "loss": -0.004506874829530716,
    "loss_tokens": 1124,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.1375408172607422,
    "outcome": "correct",
    "prompt_weight": 7.1174377224199285e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.12722118198871613,
    "support_outside_tokens": 1,
    "task_id": "211c35a89bece9052a56bf99ebc406ef800a63fc7c9992212a9685240a27353a",
    "tokens": 1124
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.13383762538433075,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.008565031923353672,
    "loss": -0.0045072282664477825,
    "loss_tokens": 2823,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.13383762538433075,
    "outcome": "correct",
    "prompt_weight": 2.8338646829613884e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.09992150962352753,
    "support_outside_tokens": 17,
    "task_id": "0b407bb90ca657a02eeeefff97e3ca30ed492a5fd569d3e562bc8ef24ff5decc",
    "tokens": 2823
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.09652648866176605,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.009431873448193073,
    "loss": -0.004507150035351515,
    "loss_tokens": 1108,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.09652648866176605,
    "outcome": "correct",
    "prompt_weight": 7.220216606498195e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.09119811654090881,
    "support_outside_tokens": 0,
    "task_id": "0545c87932bc03d0afccf9ed2ed02a227e35aa480d17c17a5c303b73943c2ebd",
    "tokens": 1108
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.2135028839111328,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.09705086052417755,
    "loss": -0.004499250091612339,
    "loss_tokens": 219,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.2135028839111328,
    "outcome": "correct",
    "prompt_weight": 3.6529680365296805e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.1760682314634323,
    "support_outside_tokens": 1,
    "task_id": "3a9c9c2393285164d3ec1564e3b583bdcce77d2db293864744b3aa29e05ba3da",
    "tokens": 219
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.12182008475065231,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.012115816585719585,
    "loss": -0.004506907891482115,
    "loss_tokens": 2979,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.12182008475065231,
    "outcome": "correct",
    "prompt_weight": 2.685464921114468e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.11294139176607132,
    "support_outside_tokens": 2,
    "task_id": "41c6041825e6613a8f26a91de7ccaa69f3e66197bdfe52b5c564b760459106dc",
    "tokens": 2979
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.11857951432466507,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.011356303468346596,
    "loss": -0.004506976343691349,
    "loss_tokens": 1450,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.11857951432466507,
    "outcome": "correct",
    "prompt_weight": 5.517241379310345e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.11182501167058945,
    "support_outside_tokens": 0,
    "task_id": "f49dcac52bfecdf4236177cec61cfeff17f0788bdf286e7d80c5893cb94fd043",
    "tokens": 1450
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.06275583058595657,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.011771188117563725,
    "loss": -0.004506939090788364,
    "loss_tokens": 1269,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.06275583058595657,
    "outcome": "correct",
    "prompt_weight": 6.304176516942475e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.056673597544431686,
    "support_outside_tokens": 1,
    "task_id": "a5609a80782de7a379bcb404ff62841e3e9e33819bdb028f5ed3e1b96e7733af",
    "tokens": 1269
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.06086800619959831,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.019904086366295815,
    "loss": -0.004506205674260855,
    "loss_tokens": 460,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.06086800619959831,
    "outcome": "correct",
    "prompt_weight": 1.7391304347826085e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.058455534279346466,
    "support_outside_tokens": 0,
    "task_id": "31a27a7b19d9daf0248fa3d6cd2d69fe535ed8719ec9ce4d3874e637ed619363",
    "tokens": 460
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.13434094190597534,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.038547683507204056,
    "loss": -0.004504525102674961,
    "loss_tokens": 271,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.13434094190597534,
    "outcome": "correct",
    "prompt_weight": 2.952029520295203e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.12940239906311035,
    "support_outside_tokens": 0,
    "task_id": "a5c1a99dea52352c5e56a5b9d9891d71dd94b051208fff78df104277a66c6706",
    "tokens": 271
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.11160969734191895,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.01582544483244419,
    "loss": -0.0045065730810165405,
    "loss_tokens": 664,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.11160969734191895,
    "outcome": "correct",
    "prompt_weight": 1.2048192771084337e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.10004835575819016,
    "support_outside_tokens": 1,
    "task_id": "0bf10d5d4f671583dbc6c168a304b883db92684e8cfa7105fb9603766c332960",
    "tokens": 664
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.0682191401720047,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.007070118561387062,
    "loss": -0.004507362376898527,
    "loss_tokens": 3469,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.0682191401720047,
    "outcome": "correct",
    "prompt_weight": 2.306140098010954e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.06232132390141487,
    "support_outside_tokens": 2,
    "task_id": "13157aaeda469fde94df9ea04628c7e129e8fe7a21733fa6f53beac4b30dcf88",
    "tokens": 3469
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.16280964016914368,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.017475508153438568,
    "loss": -0.004506424535065889,
    "loss_tokens": 311,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.16280964016914368,
    "outcome": "correct",
    "prompt_weight": 2.572347266881029e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.15427561104297638,
    "support_outside_tokens": 0,
    "task_id": "7284c6fe04f9fcc35cd25bac713b5f127d011c21be703cdeeb9a6e590ff5384a",
    "tokens": 311
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.07108867913484573,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.012678242288529873,
    "loss": -0.004506857134401798,
    "loss_tokens": 456,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.07108867913484573,
    "outcome": "correct",
    "prompt_weight": 1.7543859649122806e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.06786749511957169,
    "support_outside_tokens": 0,
    "task_id": "76134b1a83640bb29536a3a0184943a5c42483780b8ead987c29c59e6e371ee1",
    "tokens": 456
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.18155305087566376,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.05216428264975548,
    "loss": -0.004503296688199043,
    "loss_tokens": 203,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.18155305087566376,
    "outcome": "correct",
    "prompt_weight": 3.940886699507389e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.17490465939044952,
    "support_outside_tokens": 0,
    "task_id": "d35718c793470eeb8f2c61109b019cba75bb70c623b284fcc1cc48351570819f",
    "tokens": 203
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.07129190117120743,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.011329693719744682,
    "loss": -0.004506978671997786,
    "loss_tokens": 799,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.07129190117120743,
    "outcome": "correct",
    "prompt_weight": 1.0012515644555695e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.06793646514415741,
    "support_outside_tokens": 0,
    "task_id": "ef9fc35202da7ab179894c4fc6253d848166e49461d2ca20525714274ebb90ca",
    "tokens": 799
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.1561650186777115,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.07939301431179047,
    "loss": -0.004500841721892357,
    "loss_tokens": 189,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.1561650186777115,
    "outcome": "correct",
    "prompt_weight": 4.232804232804233e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.1502460539340973,
    "support_outside_tokens": 0,
    "task_id": "9fd5af81c675fe33c89e89e8d235e9fcd7aff2775f881558d262a02fbbfbedb5",
    "tokens": 189
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.09552893042564392,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.025624992325901985,
    "loss": -0.004505689721554518,
    "loss_tokens": 331,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.09552893042564392,
    "outcome": "correct",
    "prompt_weight": 2.4169184290030212e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.09107400476932526,
    "support_outside_tokens": 0,
    "task_id": "a607568da5e76f24b85d930f0b6ade1b4b0ca72c2b73c5845492699b9cbb48ed",
    "tokens": 331
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.10673835873603821,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.011026963591575623,
    "loss": -0.004507006146013737,
    "loss_tokens": 2430,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.10673835873603821,
    "outcome": "correct",
    "prompt_weight": 3.292181069958848e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.09703502058982849,
    "support_outside_tokens": 3,
    "task_id": "58492be693417fc6b46aa5a8f58e8ba385359e7362d029145fdeea0f859fb14d",
    "tokens": 2430
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.0843220204114914,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.014848592691123486,
    "loss": -0.004506661556661129,
    "loss_tokens": 551,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.0843220204114914,
    "outcome": "correct",
    "prompt_weight": 1.4519056261343013e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.08059926331043243,
    "support_outside_tokens": 0,
    "task_id": "9bb2c40065b39cb6ba35bd202b8f40b821241e5e318d6478038b743b398401d4",
    "tokens": 551
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.26175805926322937,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.026127800345420837,
    "loss": -0.004505644086748362,
    "loss_tokens": 396,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.26175805926322937,
    "outcome": "correct",
    "prompt_weight": 2.0202020202020203e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.19676139950752258,
    "support_outside_tokens": 5,
    "task_id": "5c1e708d228de272a72281f5d2d00f37bf4573000ab8fecff240d1c4316127e9",
    "tokens": 396
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.14863738417625427,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.05835219472646713,
    "loss": -0.004502738825976849,
    "loss_tokens": 245,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.14863738417625427,
    "outcome": "correct",
    "prompt_weight": 3.265306122448979e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.11477772891521454,
    "support_outside_tokens": 1,
    "task_id": "2e497d1c665174f37a39c52e1bc2b4910930439654c7f13cbeac8a508b106a78",
    "tokens": 245
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.06919591873884201,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.007398013025522232,
    "loss": -0.004507333040237427,
    "loss_tokens": 1443,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.06919591873884201,
    "outcome": "correct",
    "prompt_weight": 5.544005544005544e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.06591453403234482,
    "support_outside_tokens": 0,
    "task_id": "68692eaa82eff5079d33d75e4a4ef06adcf3c17af4d2419ac46952046cbfd52e",
    "tokens": 1443
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.11244138330221176,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.03705760836601257,
    "loss": -0.004504659213125706,
    "loss_tokens": 386,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.11244138330221176,
    "outcome": "correct",
    "prompt_weight": 2.072538860103627e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.09724315255880356,
    "support_outside_tokens": 1,
    "task_id": "483d8493e9a1357024ab8fb67fe1b8e54171285fb53d844ba4a17f7d795733ed",
    "tokens": 386
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.05774273723363876,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.011176482774317265,
    "loss": -0.004506992641836405,
    "loss_tokens": 1049,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.05774273723363876,
    "outcome": "correct",
    "prompt_weight": 7.626310772163966e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.04794608801603317,
    "support_outside_tokens": 2,
    "task_id": "8b321da461edefde469931f945899bea6bca0fb97c14becb05ed742a96ec5932",
    "tokens": 1049
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.08752410858869553,
    "advantage": 0.9970346689224243,
    "derail_index": null,
    "k3_kl_mean": 0.0062973881140351295,
    "loss": -0.004494064953178167,
    "loss_tokens": 6013,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.08752410858869553,
    "outcome": "correct",
    "prompt_weight": 1.3304506901712955e-06,
    "reward": 0.9970346715328468,
    "stop": "eos",
    "support_logq_mean": -0.08252309262752533,
    "support_outside_tokens": 1,
    "task_id": "fff4b30422ee0d3f21ef8d031119b89801ac5040797fa7637d11189ae508e9c3",
    "tokens": 6013
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.08909084647893906,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.016384968534111977,
    "loss": -0.004506522789597511,
    "loss_tokens": 612,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.08909084647893906,
    "outcome": "correct",
    "prompt_weight": 1.30718954248366e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.07258917391300201,
    "support_outside_tokens": 2,
    "task_id": "aff3cebcdd471ce6cc3923f46651c3e60a1c60a659e3ee6152c03ef0db6574df",
    "tokens": 612
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.10445711016654968,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.06682037562131882,
    "loss": -0.004501975607126951,
    "loss_tokens": 250,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.10445711016654968,
    "outcome": "correct",
    "prompt_weight": 3.2e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.07623329013586044,
    "support_outside_tokens": 1,
    "task_id": "d71d5be03fa429509ba448dbd9db6d8ec2a77646c68fc142efd61ef4b51703f3",
    "tokens": 250
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.11141061782836914,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.01207207515835762,
    "loss": -0.004506911616772413,
    "loss_tokens": 615,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.11141061782836914,
    "outcome": "correct",
    "prompt_weight": 1.3008130081300813e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.1063399612903595,
    "support_outside_tokens": 0,
    "task_id": "cc81149f4ae6c4a0d8e8276388b86b3ef48cc96a194f9cbfa456c6e62b9018dc",
    "tokens": 615
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.10960245132446289,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.012091646902263165,
    "loss": -0.004506910219788551,
    "loss_tokens": 1009,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.10960245132446289,
    "outcome": "correct",
    "prompt_weight": 7.928642220019822e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.10343893617391586,
    "support_outside_tokens": 0,
    "task_id": "cccbcd368f04fd6849292d3edbd9498b41702a38ef0c44e35ccd386e96017769",
    "tokens": 1009
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.07710941135883331,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.010668297298252583,
    "loss": -0.004507038276642561,
    "loss_tokens": 916,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.07710941135883331,
    "outcome": "correct",
    "prompt_weight": 8.733624454148471e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.068833127617836,
    "support_outside_tokens": 1,
    "task_id": "36de2ba9c7a80873493f46295f9716a0f84ab342fb9de23b5c0256cceb413969",
    "tokens": 916
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.051084112375974655,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.008932196535170078,
    "loss": -0.004507194738835096,
    "loss_tokens": 619,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.051084112375974655,
    "outcome": "correct",
    "prompt_weight": 1.2924071082390954e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.04819454625248909,
    "support_outside_tokens": 0,
    "task_id": "6012373932147e4abaae370d3d34472bf7b49177d95826a4687b3fd1985921c5",
    "tokens": 619
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.061476778239011765,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.004578016698360443,
    "loss": -0.004507587756961584,
    "loss_tokens": 4959,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.061476778239011765,
    "outcome": "correct",
    "prompt_weight": 1.613228473482557e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.055190686136484146,
    "support_outside_tokens": 4,
    "task_id": "8ce404030fc292408b1610d7dc822c1da42436a17c30d52042af714726b2bf5d",
    "tokens": 4959
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.08705132454633713,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.011316749267280102,
    "loss": -0.0045069800689816475,
    "loss_tokens": 2075,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.08705132454633713,
    "outcome": "correct",
    "prompt_weight": 3.855421686746988e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.080863356590271,
    "support_outside_tokens": 1,
    "task_id": "c8d0735cf958b0914d1b0f795116684dde65ed0f73f5236bc2db83e42ce16129",
    "tokens": 2075
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.12346596270799637,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.03107227198779583,
    "loss": -0.004505198914557695,
    "loss_tokens": 374,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.12346596270799637,
    "outcome": "correct",
    "prompt_weight": 2.1390374331550803e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.0970785915851593,
    "support_outside_tokens": 2,
    "task_id": "86cfb1c502c1d7f11de521391a9244c96ecc09c13b218a0063a7fb32650eaa22",
    "tokens": 374
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.14936339855194092,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.03703150898218155,
    "loss": -0.004504661541432142,
    "loss_tokens": 256,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.14936339855194092,
    "outcome": "correct",
    "prompt_weight": 3.125e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.12829993665218353,
    "support_outside_tokens": 1,
    "task_id": "24c8a25321edce466f3e3f5b40fe7e00739a039d434220810ae5a8d40ace4589",
    "tokens": 256
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.12247734516859055,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.01236799918115139,
    "loss": -0.004506885074079037,
    "loss_tokens": 1078,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.12247734516859055,
    "outcome": "correct",
    "prompt_weight": 7.421150278293136e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.11355019360780716,
    "support_outside_tokens": 1,
    "task_id": "e87b6c1304d795f10e08fb95c7eb23f2f9aaea6c268fd81e339a0bef1479e2b0",
    "tokens": 1078
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.08903636783361435,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.011218855157494545,
    "loss": -0.004506988916546106,
    "loss_tokens": 963,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.08903636783361435,
    "outcome": "correct",
    "prompt_weight": 8.307372793354102e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.08505879342556,
    "support_outside_tokens": 0,
    "task_id": "ba6d93ec1099be804f5fbbd88423df4806d321248eb7c5920405f1b18d29262b",
    "tokens": 963
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.07819629460573196,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.009417656809091568,
    "loss": -0.004507150966674089,
    "loss_tokens": 835,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.07819629460573196,
    "outcome": "correct",
    "prompt_weight": 9.580838323353293e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.0707397386431694,
    "support_outside_tokens": 1,
    "task_id": "293a02af5e505e4417db029c96706265de72dde0246b86b6e854cf36451b28f6",
    "tokens": 835
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.10732913017272949,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.007987463846802711,
    "loss": -0.004507279954850674,
    "loss_tokens": 1100,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.10732913017272949,
    "outcome": "correct",
    "prompt_weight": 7.272727272727272e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.10220502316951752,
    "support_outside_tokens": 0,
    "task_id": "4196968a801934f0e876666b164e7afa49222d8f948d2921de1e72d9723835b1",
    "tokens": 1100
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.0979313850402832,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.02226446568965912,
    "loss": -0.004505992867052555,
    "loss_tokens": 533,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.0979313850402832,
    "outcome": "correct",
    "prompt_weight": 1.50093808630394e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.07945860922336578,
    "support_outside_tokens": 2,
    "task_id": "aedf1c445322c33c1fec5b93c17f36b951b0d65ebcf95d38a53f669d0ffecf96",
    "tokens": 533
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.07079989463090897,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.02137516252696514,
    "loss": -0.004506072960793972,
    "loss_tokens": 330,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.07079989463090897,
    "outcome": "correct",
    "prompt_weight": 2.4242424242424244e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.06641216576099396,
    "support_outside_tokens": 0,
    "task_id": "ccf7c349ebd12019071e6d746f0e981fb57b52b15b130c259057da060d13a0f9",
    "tokens": 330
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.1418890804052353,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.04065888375043869,
    "loss": -0.004504334181547165,
    "loss_tokens": 341,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.1418890804052353,
    "outcome": "correct",
    "prompt_weight": 2.346041055718475e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.13641202449798584,
    "support_outside_tokens": 0,
    "task_id": "d26d24e9df8d6e90f06e0379fd2b15973c134528b7362dfe4112ae2a16352728",
    "tokens": 341
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.08213947713375092,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.014083806425333023,
    "loss": -0.004506730008870363,
    "loss_tokens": 369,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.08213947713375092,
    "outcome": "correct",
    "prompt_weight": 2.168021680216802e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.07809658348560333,
    "support_outside_tokens": 0,
    "task_id": "f8f99219a18a371df0b5ece9ab3f71e6ca34801ddf02b1cb24d072056ccb031d",
    "tokens": 369
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.195102721452713,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.06460987031459808,
    "loss": -0.004502174910157919,
    "loss_tokens": 255,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.195102721452713,
    "outcome": "correct",
    "prompt_weight": 3.137254901960784e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.1477418839931488,
    "support_outside_tokens": 3,
    "task_id": "6f2ebfdc856b1bde94ef369c23286b3bdb3838c52ffdd3821ebe66e4de746306",
    "tokens": 255
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.09076141566038132,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.03436625748872757,
    "loss": -0.004504901822656393,
    "loss_tokens": 483,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.09076141566038132,
    "outcome": "correct",
    "prompt_weight": 1.6563146997929607e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.07940264046192169,
    "support_outside_tokens": 1,
    "task_id": "dbd0adfa86e581b5b2db01ccfe1159be08509f589828c7dff49899506cd98363",
    "tokens": 483
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.09321478754281998,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.007823866792023182,
    "loss": -0.0045072948560118675,
    "loss_tokens": 2713,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.09321478754281998,
    "outcome": "correct",
    "prompt_weight": 2.948765204570586e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.07876436412334442,
    "support_outside_tokens": 5,
    "task_id": "2ba377a78a35d87e5e2274ee864038f5084b1186424c901080b7ab99a24edfc4",
    "tokens": 2713
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.1037849709391594,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.010073334909975529,
    "loss": -0.004507091827690601,
    "loss_tokens": 1246,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.1037849709391594,
    "outcome": "correct",
    "prompt_weight": 6.420545746388443e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.09903591871261597,
    "support_outside_tokens": 0,
    "task_id": "ef1de3af0a039a53978b7ededad12fa3389d99693c4753078bbf5967bbc90517",
    "tokens": 1246
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.06544176489114761,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.006568777374923229,
    "loss": -0.004507407546043396,
    "loss_tokens": 4937,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.06544176489114761,
    "outcome": "correct",
    "prompt_weight": 1.6204172574437917e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.06190144270658493,
    "support_outside_tokens": 0,
    "task_id": "10f0ab3e4a9e8a16cd6fc962cba3bb2ca546cf9896f85fcc45fa8787c2dd1aac",
    "tokens": 4937
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.08269284665584564,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.035862307995557785,
    "loss": -0.004504766780883074,
    "loss_tokens": 243,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.08269284665584564,
    "outcome": "correct",
    "prompt_weight": 3.292181069958848e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.04382029175758362,
    "support_outside_tokens": 2,
    "task_id": "643c5a6647d87bdd3238dcfdbacd5b75a75f0f955b3fdef9863b7f86aedeeb7f",
    "tokens": 243
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.108317069709301,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.039015304297208786,
    "loss": -0.004504482261836529,
    "loss_tokens": 549,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.108317069709301,
    "outcome": "correct",
    "prompt_weight": 1.4571948998178506e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.08056489378213882,
    "support_outside_tokens": 3,
    "task_id": "6755f115c622b906a4621a29b77e4da954a99385a070bb934f0340d6de00c26c",
    "tokens": 549
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.09181354939937592,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.011484496295452118,
    "loss": -0.004506964702159166,
    "loss_tokens": 676,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.09181354939937592,
    "outcome": "correct",
    "prompt_weight": 1.1834319526627219e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.08812834322452545,
    "support_outside_tokens": 0,
    "task_id": "650ec37d633c3738648a30bf36196ed245fa82b9fd49d6497577a9b2847721ac",
    "tokens": 676
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.051028475165367126,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.01218881644308567,
    "loss": -0.0045069013722240925,
    "loss_tokens": 385,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.051028475165367126,
    "outcome": "correct",
    "prompt_weight": 2.0779220779220778e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.037085603922605515,
    "support_outside_tokens": 1,
    "task_id": "daf78c921424073b354749eed93a64d4a4e96ac3373795e2b520aba38dbbf85c",
    "tokens": 385
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.09175072610378265,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.017417730763554573,
    "loss": -0.004506430123001337,
    "loss_tokens": 258,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.09175072610378265,
    "outcome": "correct",
    "prompt_weight": 3.1007751937984497e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.08537465333938599,
    "support_outside_tokens": 0,
    "task_id": "bb2c845b2a0f61a62b951d3a90aadf3a674d9f8f38f455e9c411853b73fbe279",
    "tokens": 258
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.07913440465927124,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.008397732861340046,
    "loss": -0.004507242701947689,
    "loss_tokens": 2509,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.07913440465927124,
    "outcome": "correct",
    "prompt_weight": 3.1885213232363493e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.06677784025669098,
    "support_outside_tokens": 2,
    "task_id": "c4a597e17c475ad6ff2d28fae2e9813f48164f1ac7ed0b9eecdeececacc541e8",
    "tokens": 2509
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.14588628709316254,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.02932082489132881,
    "loss": -0.004505356773734093,
    "loss_tokens": 267,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.14588628709316254,
    "outcome": "correct",
    "prompt_weight": 2.99625468164794e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.13938230276107788,
    "support_outside_tokens": 0,
    "task_id": "6c976b8c52be79bc4335f0f926fce070f73eac35ffa6b076b2ebcf66fa70a4d1",
    "tokens": 267
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.14397640526294708,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.023551184684038162,
    "loss": -0.0045058769173920155,
    "loss_tokens": 693,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.14397640526294708,
    "outcome": "correct",
    "prompt_weight": 1.1544011544011544e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.13243451714515686,
    "support_outside_tokens": 1,
    "task_id": "65a3f5d066f90984a8a2124573a04a364cd3d23c6a9dd170d78adb2ae831872a",
    "tokens": 693
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.1368124932050705,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.01473279669880867,
    "loss": -0.00450667180120945,
    "loss_tokens": 532,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.1368124932050705,
    "outcome": "correct",
    "prompt_weight": 1.5037593984962406e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.1289283037185669,
    "support_outside_tokens": 0,
    "task_id": "0626c1b6225cdd3cd14c636d04112c3d66be2201b3d32041143168e0e29dbd5d",
    "tokens": 532
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.1515173316001892,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.0184746403247118,
    "loss": -0.004506334196776152,
    "loss_tokens": 357,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.1515173316001892,
    "outcome": "correct",
    "prompt_weight": 2.2408963585434174e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.1450902670621872,
    "support_outside_tokens": 0,
    "task_id": "17f7ae2d3962515b5a7922304db41d0bc1f49af3b3ae7941990b4e5e3cd6dbb8",
    "tokens": 357
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.19346678256988525,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.11688155680894852,
    "loss": -0.004497461952269077,
    "loss_tokens": 81,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.19346678256988525,
    "outcome": "correct",
    "prompt_weight": 9.876543209876543e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.18869727849960327,
    "support_outside_tokens": 0,
    "task_id": "51d5377c3168c944aee9eb97ab9c2ae843d2d6e4bdd3e84b7fc92f6f41bf651b",
    "tokens": 81
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.11168628931045532,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.08355391025543213,
    "loss": -0.0045004673302173615,
    "loss_tokens": 472,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.11168628931045532,
    "outcome": "correct",
    "prompt_weight": 1.694915254237288e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.09391269832849503,
    "support_outside_tokens": 2,
    "task_id": "2165e07714d58bac16c0171765f19f2a8323618593e1ee8cbfb17bea12a1eee0",
    "tokens": 472
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.08202150464057922,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.01020658016204834,
    "loss": -0.004507080186158419,
    "loss_tokens": 1595,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.08202150464057922,
    "outcome": "correct",
    "prompt_weight": 5.015673981191222e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.07522454112768173,
    "support_outside_tokens": 1,
    "task_id": "9f2ad5f00c6a266097e581c73f4d5aca59441c68fe1113ca2ec391fcedf68be1",
    "tokens": 1595
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.06460180878639221,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.007967755198478699,
    "loss": -0.004507281817495823,
    "loss_tokens": 3118,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.06460180878639221,
    "outcome": "correct",
    "prompt_weight": 2.5657472738935216e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.05614197999238968,
    "support_outside_tokens": 3,
    "task_id": "6cdff1d90dfd539087d2776e4a33406173e899e38b3ccde649493caa9369430a",
    "tokens": 3118
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.07251318544149399,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.018171463161706924,
    "loss": -0.004506361670792103,
    "loss_tokens": 542,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.07251318544149399,
    "outcome": "correct",
    "prompt_weight": 1.4760147601476015e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.06894289702177048,
    "support_outside_tokens": 0,
    "task_id": "9e7e2bda5b28220099edc9d76b912f69346dbbfe8dbae4255f1f83463b5e848e",
    "tokens": 542
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.09657546132802963,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.007168231997638941,
    "loss": -0.004507353994995356,
    "loss_tokens": 3141,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.09657546132802963,
    "outcome": "correct",
    "prompt_weight": 2.5469595670168736e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.0911003053188324,
    "support_outside_tokens": 0,
    "task_id": "557e90483802b5eabc55d70b4f5e0852f2599b3b11f32abced72762ceb62f59c",
    "tokens": 3141
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.09411805123090744,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.009067097678780556,
    "loss": -0.004507182631641626,
    "loss_tokens": 1164,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.09411805123090744,
    "outcome": "correct",
    "prompt_weight": 6.8728522336769755e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.08941905945539474,
    "support_outside_tokens": 0,
    "task_id": "d440306a56134e5dc6f4ef006714473385f99393b134b1e2fdeedb8fdaad808d",
    "tokens": 1164
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.09558916836977005,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.024174286052584648,
    "loss": -0.004505820572376251,
    "loss_tokens": 756,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.09558916836977005,
    "outcome": "correct",
    "prompt_weight": 1.0582010582010582e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.09096258878707886,
    "support_outside_tokens": 0,
    "task_id": "1584165c54c7d8497377725e08bc49384270ca4dde1aed095f4f251ecf0295d0",
    "tokens": 756
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.07824530452489853,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.010331248864531517,
    "loss": -0.004507068544626236,
    "loss_tokens": 957,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.07824530452489853,
    "outcome": "correct",
    "prompt_weight": 8.359456635318703e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.07419051229953766,
    "support_outside_tokens": 0,
    "task_id": "d2cb473c5c44a00a6811bb6c365582a0a647e9616c54389d0e60bf79b45b439e",
    "tokens": 957
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.18675990402698517,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.05117840692400932,
    "loss": -0.004503386095166206,
    "loss_tokens": 267,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.18675990402698517,
    "outcome": "correct",
    "prompt_weight": 2.99625468164794e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.1798463761806488,
    "support_outside_tokens": 0,
    "task_id": "a88e80fe3a4c4aa396c849b2aaa0d91fc88b4c45fa449361c2ba5d6f0b5ef141",
    "tokens": 267
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.05804389715194702,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.0063316961750388145,
    "loss": -0.004507428966462612,
    "loss_tokens": 1756,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.05804389715194702,
    "outcome": "correct",
    "prompt_weight": 4.555808656036447e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.05344783514738083,
    "support_outside_tokens": 1,
    "task_id": "2ae8c1c65bc0daeee9421e32938642d8088bb019207d770f2646d4f06ffe6ed8",
    "tokens": 1756
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.15729425847530365,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.0179457850754261,
    "loss": -0.004506382159888744,
    "loss_tokens": 2412,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.15729425847530365,
    "outcome": "correct",
    "prompt_weight": 3.316749585406302e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.14618127048015594,
    "support_outside_tokens": 2,
    "task_id": "f94a348185e0992e971ad899259e236fa70c174e04ddec4f503e19362ff26b56",
    "tokens": 2412
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.08393126726150513,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.017948998138308525,
    "loss": -0.004506381694227457,
    "loss_tokens": 539,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.08393126726150513,
    "outcome": "correct",
    "prompt_weight": 1.4842300556586271e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.0684124007821083,
    "support_outside_tokens": 2,
    "task_id": "dbebf92f83c2e084cf1c9cd1ef335cbb6f8f27bfb4d814909a1f93930f56191c",
    "tokens": 539
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.1328120231628418,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.02035210095345974,
    "loss": -0.004506165627390146,
    "loss_tokens": 353,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.1328120231628418,
    "outcome": "correct",
    "prompt_weight": 2.26628895184136e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.12602020800113678,
    "support_outside_tokens": 0,
    "task_id": "e9251c941453d224e126f90ac6a5ddff61c8990d807d86286efd8aaa98fafbcf",
    "tokens": 353
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.1282133013010025,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.010637675411999226,
    "loss": -0.0045070406049489975,
    "loss_tokens": 1888,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.1282133013010025,
    "outcome": "correct",
    "prompt_weight": 4.23728813559322e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.12171699106693268,
    "support_outside_tokens": 0,
    "task_id": "3a62740a8d167ee4152caa59cd76836641024499490787bd958fcc559cc9ade4",
    "tokens": 1888
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.0866834968328476,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.012148566544055939,
    "loss": -0.004506904631853104,
    "loss_tokens": 554,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.0866834968328476,
    "outcome": "correct",
    "prompt_weight": 1.444043321299639e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.08190783113241196,
    "support_outside_tokens": 0,
    "task_id": "d376afc499a717895ee9efa5b996df5447627fcf7b1f3bb7d591122f35831dbd",
    "tokens": 554
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.06505529582500458,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.0074540539644658566,
    "loss": -0.004507327917963266,
    "loss_tokens": 1317,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.06505529582500458,
    "outcome": "correct",
    "prompt_weight": 6.074411541381929e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.062213100492954254,
    "support_outside_tokens": 0,
    "task_id": "dd4e3ba539a66ab4e9ff205c957345e2abb6ea7579e070a98c8d83790c488938",
    "tokens": 1317
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.06214437633752823,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.0084462258964777,
    "loss": -0.004507238510996103,
    "loss_tokens": 1691,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.06214437633752823,
    "outcome": "correct",
    "prompt_weight": 4.7309284447072735e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.059140972793102264,
    "support_outside_tokens": 0,
    "task_id": "6f173dc5463b844a217a788360fb042b249bbf537bd15a5aae5af6364de8a323",
    "tokens": 1691
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.09859421104192734,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.020577676594257355,
    "loss": -0.004506145138293505,
    "loss_tokens": 566,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.09859421104192734,
    "outcome": "correct",
    "prompt_weight": 1.4134275618374558e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.09404505044221878,
    "support_outside_tokens": 0,
    "task_id": "17af830b7e312c12fc7dba96357ed7f566373f1dfaf64a8fe58e7233c9328c0f",
    "tokens": 566
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.10008860379457474,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.014502829872071743,
    "loss": -0.004506692290306091,
    "loss_tokens": 660,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.10008860379457474,
    "outcome": "correct",
    "prompt_weight": 1.2121212121212122e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.09015195816755295,
    "support_outside_tokens": 1,
    "task_id": "1467450fa543c5b3d229b9799e56029649bf63ea18fb8282c3d1e5af7660befd",
    "tokens": 660
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.05609476566314697,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.006894075311720371,
    "loss": -0.004507378675043583,
    "loss_tokens": 2370,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.05609476566314697,
    "outcome": "correct",
    "prompt_weight": 3.3755274261603377e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.05202307179570198,
    "support_outside_tokens": 1,
    "task_id": "f17fb07b7134ac7ba12e82e3b1bfc2186dae25a22865125c95e2f5425d0c65bf",
    "tokens": 2370
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.08530446141958237,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.025103634223341942,
    "loss": -0.004505736753344536,
    "loss_tokens": 161,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.08530446141958237,
    "outcome": "correct",
    "prompt_weight": 4.968944099378882e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.0783766433596611,
    "support_outside_tokens": 0,
    "task_id": "c1c57e5bfa04c5711bc183176883f8b756a7126f02f453e9e87e4a9ffa3750dd",
    "tokens": 161
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.08265361934900284,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.008899577893316746,
    "loss": -0.00450719753280282,
    "loss_tokens": 928,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.08265361934900284,
    "outcome": "correct",
    "prompt_weight": 8.620689655172414e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.07861949503421783,
    "support_outside_tokens": 0,
    "task_id": "254cb3d02e78185943668f2b6a286861b0331ad248ef70570b55c35ad3199e73",
    "tokens": 928
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.06728936731815338,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.007968765683472157,
    "loss": -0.004507281817495823,
    "loss_tokens": 532,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.06728936731815338,
    "outcome": "correct",
    "prompt_weight": 1.5037593984962406e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.06297874450683594,
    "support_outside_tokens": 0,
    "task_id": "b615bf2c9ad5c11d864e9b224a9900053583616f9198d57def21fd14f50fa123",
    "tokens": 532
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.14730055630207062,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.00925101526081562,
    "loss": -0.004507166333496571,
    "loss_tokens": 2973,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.14730055630207062,
    "outcome": "correct",
    "prompt_weight": 2.690884628321561e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.13932311534881592,
    "support_outside_tokens": 0,
    "task_id": "7f0283b76218af26e3814c16e751367935323585bf7803fb0a5cc55d2d9f1694",
    "tokens": 2973
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.11249635368585587,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.010500282049179077,
    "loss": -0.004507053177803755,
    "loss_tokens": 1146,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.11249635368585587,
    "outcome": "correct",
    "prompt_weight": 6.980802792321117e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.10119999945163727,
    "support_outside_tokens": 2,
    "task_id": "bcaf170472342f98cbcc8cbb409d488d4a29a0e9e56244150a9c5e32dcdd85a7",
    "tokens": 1146
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.08562927693128586,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.017922282218933105,
    "loss": -0.004506384022533894,
    "loss_tokens": 403,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.08562927693128586,
    "outcome": "correct",
    "prompt_weight": 1.9851116625310175e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.07363913208246231,
    "support_outside_tokens": 1,
    "task_id": "4b28ca539a18f695e188274c76b9d34ac2ef1506c6a9588ab6df30bb9597dc0d",
    "tokens": 403
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.1171383261680603,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.012323504313826561,
    "loss": -0.0045068892650306225,
    "loss_tokens": 826,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.1171383261680603,
    "outcome": "correct",
    "prompt_weight": 9.685230024213076e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.11073354631662369,
    "support_outside_tokens": 0,
    "task_id": "77eb21c9e5f6c96e6e84fd195287e45ae74abd5f71c1be4d2b6f08df06c61c47",
    "tokens": 826
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.0637982189655304,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.006197402253746986,
    "loss": -0.004507441073656082,
    "loss_tokens": 1390,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.0637982189655304,
    "outcome": "correct",
    "prompt_weight": 5.755395683453238e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.06048004329204559,
    "support_outside_tokens": 0,
    "task_id": "8152065ad27babc7aaa1d6e3e7806016050ca4cb961b56f30b522cd1f9cbc9bb",
    "tokens": 1390
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.07807647436857224,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.027058687061071396,
    "loss": -0.004505560267716646,
    "loss_tokens": 503,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.07807647436857224,
    "outcome": "correct",
    "prompt_weight": 1.5904572564612325e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.07459647208452225,
    "support_outside_tokens": 0,
    "task_id": "ed2996d75339c125ce7d2e361eb56c55d934115281b960cd983bbc73e473077b",
    "tokens": 503
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.07899908721446991,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.009367142803966999,
    "loss": -0.0045071556232869625,
    "loss_tokens": 3965,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.07899908721446991,
    "outcome": "correct",
    "prompt_weight": 2.0176544766708702e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.07322520762681961,
    "support_outside_tokens": 2,
    "task_id": "61c4299e2db414a4733510bb2833ca697fa3401ee9c7c5d37a8f112aa1306877",
    "tokens": 3965
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.07405528426170349,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.0077134124003350735,
    "loss": -0.004507304634898901,
    "loss_tokens": 1320,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.07405528426170349,
    "outcome": "correct",
    "prompt_weight": 6.060606060606061e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.07040184736251831,
    "support_outside_tokens": 0,
    "task_id": "bbcfdc7dd0700462c9261d95fb6eb1b13e24d2e1f6ecbdccaaeb171d2fc8e0c7",
    "tokens": 1320
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.06978656351566315,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.019125597551465034,
    "loss": -0.004506275523453951,
    "loss_tokens": 1209,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.06978656351566315,
    "outcome": "correct",
    "prompt_weight": 6.617038875103391e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.062045641243457794,
    "support_outside_tokens": 1,
    "task_id": "c9329623ab982826ad353e80be97e4a03e35b46598fac7585206776cab499075",
    "tokens": 1209
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.1000257283449173,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.015022924169898033,
    "loss": -0.0045066457241773605,
    "loss_tokens": 323,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.1000257283449173,
    "outcome": "correct",
    "prompt_weight": 2.4767801857585138e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.09541703015565872,
    "support_outside_tokens": 0,
    "task_id": "eb6eafd5633cd628d12d8a2fde6276e1c00eeb03342be8990848195ea11615e1",
    "tokens": 323
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.166652113199234,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.043717339634895325,
    "loss": -0.004504058510065079,
    "loss_tokens": 289,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.166652113199234,
    "outcome": "correct",
    "prompt_weight": 2.7681660899653978e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.14718255400657654,
    "support_outside_tokens": 1,
    "task_id": "842eab330f748d2be004df1eb28563440d937d2d7f33de433d8e4f52707e5613",
    "tokens": 289
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.12271906435489655,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.02744268625974655,
    "loss": -0.004505525808781385,
    "loss_tokens": 368,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.12271906435489655,
    "outcome": "correct",
    "prompt_weight": 2.173913043478261e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.11674798280000687,
    "support_outside_tokens": 0,
    "task_id": "17924edb1035bba8c5425b6dae8359ea7157a96a5bbcb123f15cf935cf4cd491",
    "tokens": 368
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.11444026976823807,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.04058044031262398,
    "loss": -0.0045043411664664745,
    "loss_tokens": 228,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.11444026976823807,
    "outcome": "correct",
    "prompt_weight": 3.508771929824561e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.10930120199918747,
    "support_outside_tokens": 0,
    "task_id": "0c056374b2f2d31fb850014daae4fe8de6d942f1dcb28507e0ca53eb548b9f45",
    "tokens": 228
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.0533951111137867,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.013705514371395111,
    "loss": -0.004506764002144337,
    "loss_tokens": 855,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.0533951111137867,
    "outcome": "correct",
    "prompt_weight": 9.35672514619883e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.04726329818367958,
    "support_outside_tokens": 1,
    "task_id": "78e518acdc813f93ef362e9b798475b237ef11caf9cb40176b9978858d4da2ec",
    "tokens": 855
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.06471603363752365,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.009954169392585754,
    "loss": -0.004507102537900209,
    "loss_tokens": 779,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.06471603363752365,
    "outcome": "correct",
    "prompt_weight": 1.0269576379974327e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.06073424220085144,
    "support_outside_tokens": 0,
    "task_id": "ca11bca64cc309a78f5c10654c2415ee8832ad65d738dbd452f28177b4c2414c",
    "tokens": 779
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.12540693581104279,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.016717789694666862,
    "loss": -0.004506492987275124,
    "loss_tokens": 1207,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.12540693581104279,
    "outcome": "correct",
    "prompt_weight": 6.628003314001657e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.11837194859981537,
    "support_outside_tokens": 0,
    "task_id": "e6f9a85d583173434203c4a4735a2ed0de713a672c5e9bcc8b9ae511fcb5e154",
    "tokens": 1207
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.0874028280377388,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.008540385402739048,
    "loss": -0.004507230129092932,
    "loss_tokens": 890,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.0874028280377388,
    "outcome": "correct",
    "prompt_weight": 8.98876404494382e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.07973630726337433,
    "support_outside_tokens": 1,
    "task_id": "f564a69990a2364e138c9bdb88ad9fcb32e59a1822b98a25d80a5fcd14d08e13",
    "tokens": 890
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.07374162971973419,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.0056207492016255856,
    "loss": -0.004507493227720261,
    "loss_tokens": 1780,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.07374162971973419,
    "outcome": "correct",
    "prompt_weight": 4.49438202247191e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.06832533329725266,
    "support_outside_tokens": 1,
    "task_id": "9ec8ed3affc97798b55692dfb71270bae78ae923aebf0d63a4525f776b9e6b15",
    "tokens": 1780
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.07007155567407608,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.007388983853161335,
    "loss": -0.004507333971560001,
    "loss_tokens": 4012,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.07007155567407608,
    "outcome": "correct",
    "prompt_weight": 1.9940179461615156e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.06204969063401222,
    "support_outside_tokens": 5,
    "task_id": "92fbed84de148788ddf911817eb6daf25679b1814af1a79ba924bef0deacedfd",
    "tokens": 4012
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.10510236024856567,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.03715619817376137,
    "loss": -0.004504650365561247,
    "loss_tokens": 182,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.10510236024856567,
    "outcome": "correct",
    "prompt_weight": 4.3956043956043955e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.08325404673814774,
    "support_outside_tokens": 1,
    "task_id": "a1fe94c2a58ebc5b6e8913c5e2e021e9b27aeec5b7c8fc4756987784e9625142",
    "tokens": 182
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.12473047524690628,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.017081454396247864,
    "loss": -0.004506459925323725,
    "loss_tokens": 755,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.12473047524690628,
    "outcome": "correct",
    "prompt_weight": 1.0596026490066225e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.11436507105827332,
    "support_outside_tokens": 1,
    "task_id": "62035b4866e962208abea8abef7eaf47a2dc11829923aadc903538c9834fd857",
    "tokens": 755
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.13861775398254395,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.02356431446969509,
    "loss": -0.0045058755204081535,
    "loss_tokens": 300,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.13861775398254395,
    "outcome": "correct",
    "prompt_weight": 2.6666666666666667e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.13360902667045593,
    "support_outside_tokens": 0,
    "task_id": "0b714011b89036aa45ceafedcf9dddf81bc237a621a33f0b088bd3305c4553ee",
    "tokens": 300
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.0718725174665451,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.006540447007864714,
    "loss": -0.00450741034001112,
    "loss_tokens": 5078,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.0718725174665451,
    "outcome": "correct",
    "prompt_weight": 1.5754233950374162e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.06756611913442612,
    "support_outside_tokens": 1,
    "task_id": "27ceb4df0f1e892f6aea9e90417536d22762bbbf38813cf325e1d9af47225bb6",
    "tokens": 5078
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.1175312027335167,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.01634470745921135,
    "loss": -0.00450652651488781,
    "loss_tokens": 428,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.1175312027335167,
    "outcome": "correct",
    "prompt_weight": 1.869158878504673e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.11145234853029251,
    "support_outside_tokens": 0,
    "task_id": "01ef4e41cf91f0d09315f14b25f66dc96ff53f7a17d4b98a6cf98ba7dc646ae8",
    "tokens": 428
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.08461452275514603,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.015004714950919151,
    "loss": -0.0045066471211612225,
    "loss_tokens": 409,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.08461452275514603,
    "outcome": "correct",
    "prompt_weight": 1.9559902200488996e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.08065194636583328,
    "support_outside_tokens": 0,
    "task_id": "0a1c14c325d1dcbfcdf01ca0eea413440ffe05d49df75e136e5eb4e964ce3456",
    "tokens": 409
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.07574383169412613,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.0061955624260008335,
    "loss": -0.0045074415393173695,
    "loss_tokens": 5156,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.07574383169412613,
    "outcome": "correct",
    "prompt_weight": 1.551590380139643e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.07171059399843216,
    "support_outside_tokens": 0,
    "task_id": "e4282e2e7376065b3e275a9273e7f9630a67c69e8f8a87353cc2caa21dc81b27",
    "tokens": 5156
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.1047888770699501,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.014058343134820461,
    "loss": -0.004506732802838087,
    "loss_tokens": 514,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.1047888770699501,
    "outcome": "correct",
    "prompt_weight": 1.556420233463035e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.09972350299358368,
    "support_outside_tokens": 0,
    "task_id": "5d897115e48c8983041eb4cc402286e58296167af49dd030d8ea136e77d60321",
    "tokens": 514
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.06749195605516434,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.01230550091713667,
    "loss": -0.004506890662014484,
    "loss_tokens": 739,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.06749195605516434,
    "outcome": "correct",
    "prompt_weight": 1.0825439783491205e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.06414404511451721,
    "support_outside_tokens": 0,
    "task_id": "03aa57f7a3ea8bfe15aae5b19727ba3d54cfa0e6a25e9c4f0e9388d89d48fa48",
    "tokens": 739
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.07718357443809509,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.010465075261890888,
    "loss": -0.004507056437432766,
    "loss_tokens": 621,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.07718357443809509,
    "outcome": "correct",
    "prompt_weight": 1.288244766505636e-05,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.07346151769161224,
    "support_outside_tokens": 0,
    "task_id": "88a6051ddc73bf38ee59ba9bccdd76a05236c6a3a0917324db478f6a10e534a7",
    "tokens": 621
   },
   {
    "actor/pg_clipfrac": 0.0,
    "actor/pg_clipfrac_lower": 0.0,
    "actor/ppo_kl": 0.0,
    "actor_logprob_mean": -0.08423292636871338,
    "advantage": 1.0,
    "derail_index": null,
    "k3_kl_mean": 0.008366507478058338,
    "loss": -0.0045072459615767,
    "loss_tokens": 1158,
    "nonce": 0,
    "old_actor_max_abs_drift": 0.0,
    "old_logprob_mean": -0.08423292636871338,
    "outcome": "correct",
    "prompt_weight": 6.90846286701209e-06,
    "reward": 1.0,
    "stop": "eos",
    "support_logq_mean": -0.0768965408205986,
    "support_outside_tokens": 1,
    "task_id": "4a9706b1d585b5e93ba72e838145cbdcaa742049af1d399bc8269c85a0693b7c",
    "tokens": 1158
   }
  ],
  "unresolved_filter": {
   "enabled": false
  },
  "unresolved_trained_rollouts": 0,
  "version": "affine-mimo-grpo-update-v1"
 },
 "turn": {
  "at": 1793453292.895935,
  "capped_rollouts": 0,
  "dropped_groups": 0,
  "file_bytes": 381381,
  "file_sha256": "e1def164a7048cf25bf89d9fb0459e45c49cec1ae546b8d2125c1239e22098f8",
  "head": "87fa740811e75e4c221258ce6cf2d8cb1db1519d69aa0a65b42cb5fddea5a499",
  "hotkey": "5C4jMYe9YuupMqKdecwJ4mhwSSL9WpoL46y88EBZwppD235z",
  "queue_position": -1,
  "queue_rule": "hotkey-registration-once",
  "reg_block": 9277452,
  "rollouts": 125,
  "skipped_not_ready_count": 1,
  "tasks": 125,
  "uid": 260
 }
}
