{
    "epoch": 2.0,
    "eval_logits/chosen": -2.702141284942627,
    "eval_logits/rejected": -2.6936912536621094,
    "eval_logps/chosen": -1604.7235107421875,
    "eval_logps/rejected": -1409.80029296875,
    "eval_loss": 0.6445866227149963,
    "eval_rewards/accuracies": 0.6289682388305664,
    "eval_rewards/chosen": -0.4217440187931061,
    "eval_rewards/margins": 0.15961241722106934,
    "eval_rewards/rejected": -0.5813564658164978,
    "eval_runtime": 221.9698,
    "eval_samples": 2000,
    "eval_samples_per_second": 9.01,
    "eval_steps_per_second": 0.284,
    "train_loss": 0.6517634629204897,
    "train_runtime": 44544.264,
    "train_samples": 61135,
    "train_samples_per_second": 2.745,
    "train_steps_per_second": 0.172
}