all_results.json

{
    "epoch": 2.994495412844037,
    "eval_logits/chosen": -0.4288872182369232,
    "eval_logits/rejected": -1.5388305187225342,
    "eval_logps/chosen": -277.0741882324219,
    "eval_logps/rejected": -182.1341094970703,
    "eval_loss": 0.010117708705365658,
    "eval_rewards/accuracies": 1.0,
    "eval_rewards/chosen": 4.345386981964111,
    "eval_rewards/margins": 9.913021087646484,
    "eval_rewards/rejected": -5.567633628845215,
    "eval_runtime": 8.9732,
    "eval_samples_per_second": 5.015,
    "eval_steps_per_second": 5.015,
    "total_flos": 7.837376281021809e+17,
    "train_loss": 0.11720214437923905,
    "train_runtime": 8069.9016,
    "train_samples_per_second": 1.62,
    "train_steps_per_second": 0.051
}