{
    "epoch": 1.0,
    "eval_logits/chosen": -1.5923750400543213,
    "eval_logits/rejected": -1.5335367918014526,
    "eval_logps/chosen": -471.41455078125,
    "eval_logps/rejected": -688.0838623046875,
    "eval_loss": 0.48019346594810486,
    "eval_rewards/accuracies": 0.7946428656578064,
    "eval_rewards/chosen": -1.9911106824874878,
    "eval_rewards/margins": 2.1531193256378174,
    "eval_rewards/rejected": -4.144230365753174,
    "eval_runtime": 45.7848,
    "eval_samples": 4461,
    "eval_samples_per_second": 97.434,
    "eval_steps_per_second": 1.529
}