all_results.json

{
    "epoch": 1.0,
    "eval_logits/chosen": -3.813626766204834,
    "eval_logits/rejected": -3.8465464115142822,
    "eval_logps/chosen": -292.01556396484375,
    "eval_logps/rejected": -185.1787109375,
    "eval_loss": 0.5876513123512268,
    "eval_rewards/accuracies": 0.6875,
    "eval_rewards/chosen": -0.18421681225299835,
    "eval_rewards/diff": -2.171558380126953,
    "eval_rewards/diff_abs": 2.208705186843872,
    "eval_rewards/rejected": -0.4189082384109497,
    "eval_rewards/student_margin": 0.23469144105911255,
    "eval_rewards/teacher_margin": 2.40625,
    "eval_runtime": 13.1071,
    "eval_samples": 1470,
    "eval_samples_per_second": 112.153,
    "eval_steps_per_second": 0.153,
    "train_loss": 0.6328112833182432,
    "train_runtime": 3006.7888,
    "train_samples": 147002,
    "train_samples_per_second": 48.89,
    "train_steps_per_second": 0.127
}