{ "epoch": 3.0, "eval_logits/chosen": -2.6452701091766357, "eval_logits/rejected": -2.628488779067993, "eval_logps/chosen": -74.79349517822266, "eval_logps/rejected": -84.08062744140625, "eval_loss": 0.8725804686546326, "eval_rewards/accuracies": 0.375, "eval_rewards/chosen": -0.1874038428068161, "eval_rewards/margins": 1.6430120468139648, "eval_rewards/rejected": -1.8304157257080078, "eval_runtime": 113.9064, "eval_samples": 2000, "eval_samples_per_second": 17.558, "eval_steps_per_second": 0.553 }