{
  "schema_version": 1,
  "model": "SlayerLab/pollock-mini-lm-125m",
  "release": "Pollock 1.4",
  "model_revision": "a2e53d9f1b690ebfbea375cf8bd3b5bdf69dece9",
  "run_date": "2026-09-03",
  "protocol": {
    "harness": "LightEval 0.6.0.dev0",
    "harness_revision": "ea46419a93fb390e8f694f7c6c64c1e684487c9d",
    "recipe": "Hugging Face SmolLM2 base evaluation",
    "recipe_revision": "d3bb683bc123a1954b289b70dd8119b51c358018",
    "recipe_file_sha256": "6899e7565b529ef2200e500fd9814c46bbdf42f9576f069d6de8a4dc2b9355b1",
    "transformers": "4.46.3",
    "torch": "2.4.1+cu121",
    "datasets": "3.1.0",
    "few_shot": 0,
    "max_samples": null,
    "batch_size": 64,
    "device": "NVIDIA GeForce RTX 3090 (CUDA device 0)",
    "dtype": "bfloat16",
    "evaluation_time_seconds": 371.367206582996
  },
  "results": {
    "mmlu_cloze": { "metric": "acc_norm", "value": 0.2671271898589944, "stderr": 0.0037340001649945943, "samples": 14042 },
    "arc_easy": { "metric": "acc_norm", "value": 0.42045454545454547, "stderr": 0.010129114278546533, "samples": 2376 },
    "arc_challenge": { "metric": "acc_norm", "value": 0.24232081911262798, "stderr": 0.012521593295800116, "samples": 1172 },
    "arc_mean": { "metric": "arithmetic_mean_of_acc_norm", "value": 0.3313876822835867, "stderr": 0.011325353787173325, "samples": 3548 },
    "piqa": { "metric": "acc_norm", "value": 0.6060935799782372, "stderr": 0.011400182224698902, "samples": 1838 },
    "hellaswag": { "metric": "acc_norm", "value": 0.29336785500896234, "stderr": 0.00454375048006577, "samples": 10042 },
    "openbookqa": { "metric": "acc_norm", "value": 0.286, "stderr": 0.020229346329177517, "samples": 500 },
    "winogrande": { "metric": "acc_norm", "value": 0.5027624309392266, "stderr": 0.014052271211616441, "samples": 1267 },
    "five_task_mean": { "metric": "unweighted_mean_of_arc_mean_piqa_hellaswag_openbookqa_winogrande", "value": 0.40392230964200254 }
  },
  "raw_result_sha256": "5579421f7ba77990acc592511701978e4d63ac61facd038f35f6152729c1a868",
  "model_weights_sha256": "2aada5b26abe9ce3b70af393d77be4ed1a799a7205b9c0dd5cfaff8bcaabce68"
}
