all_results.json

{
    "epoch": 0.99,
    "eval_logits/chosen": 0.8012984991073608,
    "eval_logits/rejected": 0.8187645673751831,
    "eval_logps/chosen": -91.72753143310547,
    "eval_logps/rejected": -94.23716735839844,
    "eval_loss": 0.00012303233961574733,
    "eval_rewards/accuracies": 0.25999999046325684,
    "eval_rewards/chosen": -1.6935970279519097e-06,
    "eval_rewards/margins": 0.0003997479216195643,
    "eval_rewards/rejected": -0.00040144147351384163,
    "eval_runtime": 274.0423,
    "eval_samples": 2000,
    "eval_samples_per_second": 7.298,
    "eval_steps_per_second": 0.456,
    "train_loss": 0.0001254068662283795,
    "train_runtime": 1151.127,
    "train_samples": 6113,
    "train_samples_per_second": 5.31,
    "train_steps_per_second": 0.083
}