{
    "epoch": 2.0,
    "eval_logits/chosen": -2.4423749446868896,
    "eval_logits/rejected": -2.424229860305786,
    "eval_logps/chosen": -76.38975524902344,
    "eval_logps/rejected": -85.61776733398438,
    "eval_loss": 0.8224331140518188,
    "eval_rewards/accuracies": 0.3432539701461792,
    "eval_rewards/chosen": -0.8205374479293823,
    "eval_rewards/margins": 1.119767189025879,
    "eval_rewards/rejected": -1.9403046369552612,
    "eval_runtime": 113.8823,
    "eval_samples": 2000,
    "eval_samples_per_second": 17.562,
    "eval_steps_per_second": 0.553,
    "total_flos": 0.0,
    "train_loss": 0.5438582648833593,
    "train_runtime": 2038.8345,
    "train_samples": 6113,
    "train_samples_per_second": 5.997,
    "train_steps_per_second": 0.094
}