all_results.json

{
    "epoch": 0.9981298423724285,
    "eval_logits/chosen": 0.19048450887203217,
    "eval_logits/rejected": 0.20416969060897827,
    "eval_logps/chosen": -159.4064483642578,
    "eval_logps/rejected": -166.90692138671875,
    "eval_loss": 0.6582256555557251,
    "eval_rewards/accuracies": 0.5927419066429138,
    "eval_rewards/chosen": -0.7111408710479736,
    "eval_rewards/margins": 0.09036804735660553,
    "eval_rewards/rejected": -0.8015088438987732,
    "eval_runtime": 180.5604,
    "eval_samples_per_second": 10.861,
    "eval_steps_per_second": 0.172
}