{
  "results": {
    "race": {
      "acc": 0.3521531100478469,
      "acc_stderr": 0.014782629897202266
    },
    "coqa": {
      "f1": 0.6149953283852732,
      "f1_stderr": 0.017671101287646335,
      "em": 0.4606666666666665,
      "em_stderr": 0.020185214533701167
    },
    "drop": {
      "em": 0.0194001677852349,
      "em_stderr": 0.0014124994962717904,
      "f1": 0.08879089765100667,
      "f1_stderr": 0.002002230574295575
    }
  },
  "versions": {
    "race": 1,
    "coqa": 1,
    "drop": 1
  },
  "config": {
    "model": "hf-causal-experimental",
    "model_args": "pretrained=bigscience/bloom-3b,use_accelerate=True",
    "num_fewshot": 0,
    "batch_size": "auto",
    "device": "cuda:0",
    "no_cache": true,
    "limit": null,
    "bootstrap_iters": 100000,
    "description_dict": {}
  }
}