{
  "_name_or_path": "trained_model",
  "architectures": [
    "RewardModel"
  ],
  "model_type": "gpt_neo",
  "torch_dtype": "float32",
  "transformers_version": "4.24.0"
}