eval_results.json

{
    "epoch": 1.0,
    "eval_logits/chosen": -0.5036042332649231,
    "eval_logits/rejected": -0.22393426299095154,
    "eval_logps/chosen": -363.8373107910156,
    "eval_logps/rejected": -1280.70751953125,
    "eval_loss": 0.05244915559887886,
    "eval_rewards/accuracies": 0.9918341636657715,
    "eval_rewards/chosen": -0.18055257201194763,
    "eval_rewards/margins": 10.881051063537598,
    "eval_rewards/rejected": -11.061604499816895,
    "eval_runtime": 1903.9644,
    "eval_samples": 3174,
    "eval_samples_per_second": 1.667,
    "eval_steps_per_second": 0.105
}