{
  "schema": "stateplay-formal-model-publication-v1",
  "repo": "lupele/rewardmodel",
  "created_at": "2026-10-22T01:22:28.623799+00:00",
  "kind": "formal_selected_reward_model",
  "version": "28befdbd2782031fe8d67975",
  "optimizer_step": 132,
  "smoke": false,
  "adapter_sha256": "81078f22d9cbdddde35f1c818d53bdf079316d88f30ca7ad63dc99e0c0e605e6",
  "base_model": "Qwen/Qwen3.5-9B",
  "base_revision": "c202236235762e1c871ad0ccb60c8ee5ba337b9a",
  "training_state": {
    "source_sha256": "70fdf96a91c23114bed6865114438acca4d1dc6289a035f869dd11b571633c58",
    "published_sha256": "c34c3a00110369181c5c63ac8d2881316c92165afe31fa89a0e2f0510f77a405",
    "metadata_only_changes": [
      "/config/reward/endpoint",
      "/config/reward_model/inference_url",
      "/config/config_hash"
    ],
    "preserved_exactly": [
      "optimizer",
      "scheduler",
      "step",
      "rng",
      "rank_rng",
      "buffers",
      "accumulation",
      "world_size",
      "total_steps",
      "backend",
      "dataset_hash",
      "smoke"
    ],
    "all_parameters_optimizer_rng_exactly_equal": true
  },
  "manual_review_pending": true
}