all_results.json

{
    "epoch": 4.0,
    "eval_logits/chosen": -2.2051267623901367,
    "eval_logits/rejected": -2.2002947330474854,
    "eval_logps/chosen": -34.13163375854492,
    "eval_logps/rejected": -37.722042083740234,
    "eval_loss": 0.9173061847686768,
    "eval_rewards/accuracies": 0.574335515499115,
    "eval_rewards/chosen": -0.08737529814243317,
    "eval_rewards/margins": 0.09750958532094955,
    "eval_rewards/rejected": -0.18488489091396332,
    "eval_runtime": 145.0469,
    "eval_samples": 343,
    "eval_samples_per_second": 2.365,
    "eval_steps_per_second": 0.296,
    "train_loss": 0.15107146371881683,
    "train_runtime": 10740.8737,
    "train_samples": 3079,
    "train_samples_per_second": 1.147,
    "train_steps_per_second": 0.143
}