{ "best_metric": null, "best_model_checkpoint": null, "epoch": 0.9984, "eval_steps": 50, "global_step": 312, "is_hyper_param_search": false, "is_local_process_zero": true, "is_world_process_zero": true, "log_history": [ { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.02890625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1456.6, "completions/mean_length": 168.94833984375, "completions/mean_terminated_length": 128.25774536132812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.016, "grad_norm": 0.03971971198916435, "learning_rate": 3.1249999999999997e-07, "loss": 0.0369, "num_tokens": 13430383.0, "reward": 0.42822265625, "reward_std": 0.3223201155662537, "rewards/accuracy_reward": 0.19130859375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.66513671875, "rewards/mean_confidence_reward": 0.0, "step": 5 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0181640625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1478.4, "completions/mean_length": 149.1314453125, "completions/mean_terminated_length": 123.49475402832032, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "epoch": 0.032, "grad_norm": 0.01608453504741192, "learning_rate": 6.249999999999999e-07, "loss": 0.0318, "num_tokens": 26914161.0, "reward": 0.477685546875, "reward_std": 0.29675968885421755, "rewards/accuracy_reward": 0.20576171875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.749609375, "rewards/mean_confidence_reward": 0.0, "step": 10 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.009765625, "completions/max_length": 1536.0, "completions/max_terminated_length": 1360.2, "completions/mean_length": 116.2447265625, "completions/mean_terminated_length": 102.26145935058594, "completions/min_length": 3.2, "completions/min_terminated_length": 3.2, "epoch": 0.048, "grad_norm": 0.024524839594960213, "learning_rate": 9.374999999999999e-07, "loss": 0.0369, "num_tokens": 40009563.0, "reward": 0.58642578125, "reward_std": 0.20457804799079896, "rewards/accuracy_reward": 0.24951171875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.92333984375, "rewards/mean_confidence_reward": 0.0, "step": 15 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.005078125, "completions/max_length": 1536.0, "completions/max_terminated_length": 1367.2, "completions/mean_length": 84.524609375, "completions/mean_terminated_length": 77.11885986328124, "completions/min_length": 6.8, "completions/min_terminated_length": 6.8, "epoch": 0.064, "grad_norm": 0.00957680307328701, "learning_rate": 1e-06, "loss": 0.0187, "num_tokens": 52649815.0, "reward": 0.65634765625, "reward_std": 0.1586296558380127, "rewards/accuracy_reward": 0.332421875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9802734375, "rewards/mean_confidence_reward": 0.0, "step": 20 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00068359375, "completions/max_length": 1287.4, "completions/max_terminated_length": 422.8, "completions/mean_length": 70.96943359375, "completions/mean_terminated_length": 69.96731414794922, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "epoch": 0.08, "grad_norm": 0.006326671689748764, "learning_rate": 1e-06, "loss": 0.0011, "num_tokens": 65166014.0, "reward": 0.69560546875, "reward_std": 0.11688852310180664, "rewards/accuracy_reward": 0.39384765625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99736328125, "rewards/mean_confidence_reward": 0.0, "step": 25 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0001953125, "completions/max_length": 843.8, "completions/max_terminated_length": 404.0, "completions/mean_length": 73.64716796875, "completions/mean_terminated_length": 73.3616714477539, "completions/min_length": 15.6, "completions/min_terminated_length": 15.6, "epoch": 0.096, "grad_norm": 0.015297948382794857, "learning_rate": 1e-06, "loss": 0.0006, "num_tokens": 77821089.0, "reward": 0.702392578125, "reward_std": 0.1077111542224884, "rewards/accuracy_reward": 0.4056640625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99912109375, "rewards/mean_confidence_reward": 0.0, "step": 30 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 1045.2, "completions/max_terminated_length": 579.6, "completions/mean_length": 77.37705078125, "completions/mean_terminated_length": 76.80772857666015, "completions/min_length": 15.4, "completions/min_terminated_length": 15.4, "epoch": 0.112, "grad_norm": 0.0034716154914349318, "learning_rate": 1e-06, "loss": 0.0027, "num_tokens": 90579222.0, "reward": 0.717138671875, "reward_std": 0.10914339721202851, "rewards/accuracy_reward": 0.4361328125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99814453125, "rewards/mean_confidence_reward": 0.0, "step": 35 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 1137.8, "completions/max_terminated_length": 429.2, "completions/mean_length": 80.57646484375, "completions/mean_terminated_length": 80.00765838623047, "completions/min_length": 20.2, "completions/min_terminated_length": 20.2, "epoch": 0.128, "grad_norm": 0.0017264570342376828, "learning_rate": 1e-06, "loss": 0.0021, "num_tokens": 103177317.0, "reward": 0.7154296875, "reward_std": 0.09922275692224503, "rewards/accuracy_reward": 0.4318359375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9990234375, "rewards/mean_confidence_reward": 0.0, "step": 40 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00029296875, "completions/max_length": 1027.8, "completions/max_terminated_length": 526.4, "completions/mean_length": 80.383203125, "completions/mean_terminated_length": 79.95665893554687, "completions/min_length": 18.2, "completions/min_terminated_length": 18.2, "epoch": 0.144, "grad_norm": 0.0022848467342555523, "learning_rate": 1e-06, "loss": 0.0006, "num_tokens": 115807193.0, "reward": 0.76044921875, "reward_std": 0.09742432534694671, "rewards/accuracy_reward": 0.5216796875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99921875, "rewards/mean_confidence_reward": 0.0, "step": 45 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00048828125, "completions/max_length": 1354.8, "completions/max_terminated_length": 391.2, "completions/mean_length": 85.2048828125, "completions/mean_terminated_length": 84.49549560546875, "completions/min_length": 29.2, "completions/min_terminated_length": 29.2, "epoch": 0.16, "grad_norm": 0.0022386491764336824, "learning_rate": 1e-06, "loss": 0.0009, "num_tokens": 128556939.0, "reward": 0.7416015625, "reward_std": 0.09557278156280517, "rewards/accuracy_reward": 0.48388671875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99931640625, "rewards/mean_confidence_reward": 0.0, "step": 50 }, { "epoch": 0.16, "eval_completions/clipped_ratio": 0.0, "eval_completions/max_length": 315.75, "eval_completions/max_terminated_length": 315.75, "eval_completions/mean_length": 93.1332836151123, "eval_completions/mean_terminated_length": 93.1332836151123, "eval_completions/min_length": 39.25, "eval_completions/min_terminated_length": 39.25, "eval_loss": 0.0, "eval_num_tokens": 128556939.0, "eval_reward": 0.6962890625, "eval_reward_std": 0.2430882230401039, "eval_rewards/accuracy_reward": 0.392578125, "eval_rewards/brier_reward": 0.0, "eval_rewards/confidence_one_or_zero": 0.0, "eval_rewards/format_reward": 1.0, "eval_rewards/mean_confidence_reward": 0.0, "eval_runtime": 17.5973, "eval_samples_per_second": 28.413, "eval_steps_per_second": 0.227, "step": 50 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00068359375, "completions/max_length": 1087.0, "completions/max_terminated_length": 626.8, "completions/mean_length": 93.46279296875, "completions/mean_terminated_length": 92.47601928710938, "completions/min_length": 30.6, "completions/min_terminated_length": 30.6, "epoch": 0.176, "grad_norm": 0.002047221176326275, "learning_rate": 1e-06, "loss": 0.0034, "num_tokens": 141607438.0, "reward": 0.7376953125, "reward_std": 0.09575197547674179, "rewards/accuracy_reward": 0.4763671875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9990234375, "rewards/mean_confidence_reward": 0.0, "step": 55 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.001171875, "completions/max_length": 1536.0, "completions/max_terminated_length": 727.4, "completions/mean_length": 94.7880859375, "completions/mean_terminated_length": 93.09821166992188, "completions/min_length": 30.8, "completions/min_terminated_length": 30.8, "epoch": 0.192, "grad_norm": 0.0016944622620940208, "learning_rate": 1e-06, "loss": 0.0038, "num_tokens": 154249204.0, "reward": 0.7494140625, "reward_std": 0.08774578422307969, "rewards/accuracy_reward": 0.5001953125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9986328125, "rewards/mean_confidence_reward": 0.0, "step": 60 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0001953125, "completions/max_length": 945.4, "completions/max_terminated_length": 503.0, "completions/mean_length": 94.95087890625, "completions/mean_terminated_length": 94.66974792480468, "completions/min_length": 33.8, "completions/min_terminated_length": 33.8, "epoch": 0.208, "grad_norm": 0.0016015366418287158, "learning_rate": 1e-06, "loss": 0.0015, "num_tokens": 167110045.0, "reward": 0.77373046875, "reward_std": 0.08670773208141327, "rewards/accuracy_reward": 0.5478515625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.999609375, "rewards/mean_confidence_reward": 0.0, "step": 65 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.001953125, "completions/max_length": 1297.4, "completions/max_terminated_length": 481.2, "completions/mean_length": 96.69267578125, "completions/mean_terminated_length": 93.87931976318359, "completions/min_length": 32.4, "completions/min_terminated_length": 32.4, "epoch": 0.224, "grad_norm": 0.001440020278096199, "learning_rate": 1e-06, "loss": 0.0051, "num_tokens": 180109682.0, "reward": 0.74384765625, "reward_std": 0.08273749649524689, "rewards/accuracy_reward": 0.48974609375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99794921875, "rewards/mean_confidence_reward": 0.0, "step": 70 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00078125, "completions/max_length": 1132.0, "completions/max_terminated_length": 595.0, "completions/mean_length": 96.2376953125, "completions/mean_terminated_length": 95.11122589111328, "completions/min_length": 32.6, "completions/min_terminated_length": 32.6, "epoch": 0.24, "grad_norm": 0.0018575695576146245, "learning_rate": 1e-06, "loss": 0.0032, "num_tokens": 193203156.0, "reward": 0.77578125, "reward_std": 0.09852926135063171, "rewards/accuracy_reward": 0.55234375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99921875, "rewards/mean_confidence_reward": 0.0, "step": 75 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 1310.2, "completions/max_terminated_length": 423.2, "completions/mean_length": 91.51962890625, "completions/mean_terminated_length": 90.95541381835938, "completions/min_length": 33.6, "completions/min_terminated_length": 33.6, "epoch": 0.256, "grad_norm": 0.0016601247480139136, "learning_rate": 1e-06, "loss": 0.0016, "num_tokens": 206051453.0, "reward": 0.76103515625, "reward_std": 0.08267472535371781, "rewards/accuracy_reward": 0.52265625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9994140625, "rewards/mean_confidence_reward": 0.0, "step": 80 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00029296875, "completions/max_length": 806.6, "completions/max_terminated_length": 347.2, "completions/mean_length": 92.82333984375, "completions/mean_terminated_length": 92.40117645263672, "completions/min_length": 33.0, "completions/min_terminated_length": 33.0, "epoch": 0.272, "grad_norm": 0.0015578230377286673, "learning_rate": 1e-06, "loss": 0.0015, "num_tokens": 218823980.0, "reward": 0.758251953125, "reward_std": 0.07986692190170289, "rewards/accuracy_reward": 0.516796875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99970703125, "rewards/mean_confidence_reward": 0.0, "step": 85 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0001953125, "completions/max_length": 742.4, "completions/max_terminated_length": 510.6, "completions/mean_length": 87.11962890625, "completions/mean_terminated_length": 86.83695526123047, "completions/min_length": 31.4, "completions/min_terminated_length": 31.4, "epoch": 0.288, "grad_norm": 0.0017029134323820472, "learning_rate": 1e-06, "loss": 0.001, "num_tokens": 231530581.0, "reward": 0.766064453125, "reward_std": 0.08002678900957108, "rewards/accuracy_reward": 0.53232421875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9998046875, "rewards/mean_confidence_reward": 0.0, "step": 90 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 1143.6, "completions/max_terminated_length": 458.8, "completions/mean_length": 87.85400390625, "completions/mean_terminated_length": 87.2877700805664, "completions/min_length": 32.6, "completions/min_terminated_length": 32.6, "epoch": 0.304, "grad_norm": 0.002200548769906163, "learning_rate": 1e-06, "loss": 0.0017, "num_tokens": 244216478.0, "reward": 0.763037109375, "reward_std": 0.07723365724086761, "rewards/accuracy_reward": 0.52646484375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.999609375, "rewards/mean_confidence_reward": 0.0, "step": 95 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00048828125, "completions/max_length": 1041.0, "completions/max_terminated_length": 307.2, "completions/mean_length": 87.27138671875, "completions/mean_terminated_length": 86.5638931274414, "completions/min_length": 36.2, "completions/min_terminated_length": 36.2, "epoch": 0.32, "grad_norm": 0.0014393636956810951, "learning_rate": 1e-06, "loss": 0.0011, "num_tokens": 257055161.0, "reward": 0.769873046875, "reward_std": 0.06611677706241607, "rewards/accuracy_reward": 0.54033203125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9994140625, "rewards/mean_confidence_reward": 0.0, "step": 100 }, { "epoch": 0.32, "eval_completions/clipped_ratio": 0.001953125, "eval_completions/max_length": 593.5, "eval_completions/max_terminated_length": 258.0, "eval_completions/mean_length": 95.18911552429199, "eval_completions/mean_terminated_length": 92.36835670471191, "eval_completions/min_length": 46.75, "eval_completions/min_terminated_length": 46.75, "eval_loss": 0.0, "eval_num_tokens": 257055161.0, "eval_reward": 0.697265625, "eval_reward_std": 0.24716106429696083, "eval_rewards/accuracy_reward": 0.396484375, "eval_rewards/brier_reward": 0.0, "eval_rewards/confidence_one_or_zero": 0.0, "eval_rewards/format_reward": 0.998046875, "eval_rewards/mean_confidence_reward": 0.0, "eval_runtime": 24.5152, "eval_samples_per_second": 20.395, "eval_steps_per_second": 0.163, "step": 100 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00048828125, "completions/max_length": 1300.4, "completions/max_terminated_length": 441.4, "completions/mean_length": 90.16708984375, "completions/mean_terminated_length": 89.46054534912109, "completions/min_length": 37.2, "completions/min_terminated_length": 37.2, "epoch": 0.336, "grad_norm": 0.0015561841428279877, "learning_rate": 1e-06, "loss": 0.0017, "num_tokens": 269557224.0, "reward": 0.775537109375, "reward_std": 0.07957550585269928, "rewards/accuracy_reward": 0.5515625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99951171875, "rewards/mean_confidence_reward": 0.0, "step": 105 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0001953125, "completions/max_length": 966.8, "completions/max_terminated_length": 503.4, "completions/mean_length": 91.57109375, "completions/mean_terminated_length": 91.28900604248047, "completions/min_length": 37.4, "completions/min_terminated_length": 37.4, "epoch": 0.352, "grad_norm": 0.0015385675942525268, "learning_rate": 1e-06, "loss": 0.0015, "num_tokens": 282611648.0, "reward": 0.749658203125, "reward_std": 0.07262765616178513, "rewards/accuracy_reward": 0.499609375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99970703125, "rewards/mean_confidence_reward": 0.0, "step": 110 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0005859375, "completions/max_length": 878.6, "completions/max_terminated_length": 388.2, "completions/mean_length": 92.1431640625, "completions/mean_terminated_length": 91.29630737304687, "completions/min_length": 36.0, "completions/min_terminated_length": 36.0, "epoch": 0.368, "grad_norm": 0.001171114738099277, "learning_rate": 1e-06, "loss": 0.0016, "num_tokens": 295476986.0, "reward": 0.756591796875, "reward_std": 0.06191314309835434, "rewards/accuracy_reward": 0.51376953125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9994140625, "rewards/mean_confidence_reward": 0.0, "step": 115 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0001953125, "completions/max_length": 655.4, "completions/max_terminated_length": 408.8, "completions/mean_length": 93.15146484375, "completions/mean_terminated_length": 92.86930847167969, "completions/min_length": 36.4, "completions/min_terminated_length": 36.4, "epoch": 0.384, "grad_norm": 0.0033440215047448874, "learning_rate": 1e-06, "loss": 0.0004, "num_tokens": 308143689.0, "reward": 0.772265625, "reward_std": 0.06511491686105728, "rewards/accuracy_reward": 0.5447265625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9998046875, "rewards/mean_confidence_reward": 0.0, "step": 120 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0, "completions/max_length": 427.8, "completions/max_terminated_length": 427.8, "completions/mean_length": 91.9720703125, "completions/mean_terminated_length": 91.9720703125, "completions/min_length": 36.0, "completions/min_terminated_length": 36.0, "epoch": 0.4, "grad_norm": 0.0016430045943707228, "learning_rate": 1e-06, "loss": 0.0005, "num_tokens": 320978251.0, "reward": 0.762353515625, "reward_std": 0.0696032926440239, "rewards/accuracy_reward": 0.5248046875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99990234375, "rewards/mean_confidence_reward": 0.0, "step": 125 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 867.8, "completions/max_terminated_length": 388.2, "completions/mean_length": 94.73369140625, "completions/mean_terminated_length": 94.17031402587891, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "epoch": 0.416, "grad_norm": 0.0013333384413272142, "learning_rate": 1e-06, "loss": 0.0015, "num_tokens": 333685828.0, "reward": 0.769775390625, "reward_std": 0.0636753872036934, "rewards/accuracy_reward": 0.53994140625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.999609375, "rewards/mean_confidence_reward": 0.0, "step": 130 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 1081.2, "completions/max_terminated_length": 420.8, "completions/mean_length": 94.04560546875, "completions/mean_terminated_length": 93.48209533691406, "completions/min_length": 42.6, "completions/min_terminated_length": 42.6, "epoch": 0.432, "grad_norm": 0.0019940051715821028, "learning_rate": 1e-06, "loss": 0.0017, "num_tokens": 346519511.0, "reward": 0.787548828125, "reward_std": 0.061328309774398806, "rewards/accuracy_reward": 0.57548828125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.999609375, "rewards/mean_confidence_reward": 0.0, "step": 135 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00029296875, "completions/max_length": 1084.4, "completions/max_terminated_length": 401.0, "completions/mean_length": 95.8486328125, "completions/mean_terminated_length": 95.42702941894531, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "epoch": 0.448, "grad_norm": 0.0024249772541224957, "learning_rate": 1e-06, "loss": 0.0015, "num_tokens": 359310121.0, "reward": 0.766064453125, "reward_std": 0.061192930489778516, "rewards/accuracy_reward": 0.53251953125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.999609375, "rewards/mean_confidence_reward": 0.0, "step": 140 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00078125, "completions/max_length": 1311.0, "completions/max_terminated_length": 497.0, "completions/mean_length": 97.95634765625, "completions/mean_terminated_length": 96.83337860107422, "completions/min_length": 41.4, "completions/min_terminated_length": 41.4, "epoch": 0.464, "grad_norm": 0.0012390408664941788, "learning_rate": 1e-06, "loss": 0.0022, "num_tokens": 372340330.0, "reward": 0.73681640625, "reward_std": 0.05486533492803573, "rewards/accuracy_reward": 0.47451171875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99912109375, "rewards/mean_confidence_reward": 0.0, "step": 145 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 1080.2, "completions/max_terminated_length": 732.2, "completions/mean_length": 98.50205078125, "completions/mean_terminated_length": 97.9401870727539, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "epoch": 0.48, "grad_norm": 0.0014968572650104761, "learning_rate": 1e-06, "loss": 0.0015, "num_tokens": 385253343.0, "reward": 0.77099609375, "reward_std": 0.07616502344608307, "rewards/accuracy_reward": 0.5423828125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.999609375, "rewards/mean_confidence_reward": 0.0, "step": 150 }, { "epoch": 0.48, "eval_completions/clipped_ratio": 0.0, "eval_completions/max_length": 291.5, "eval_completions/max_terminated_length": 291.5, "eval_completions/mean_length": 97.19706344604492, "eval_completions/mean_terminated_length": 97.19706344604492, "eval_completions/min_length": 53.75, "eval_completions/min_terminated_length": 53.75, "eval_loss": 0.0, "eval_num_tokens": 385253343.0, "eval_reward": 0.724609375, "eval_reward_std": 0.24863140657544136, "eval_rewards/accuracy_reward": 0.44921875, "eval_rewards/brier_reward": 0.0, "eval_rewards/confidence_one_or_zero": 0.0, "eval_rewards/format_reward": 1.0, "eval_rewards/mean_confidence_reward": 0.0, "eval_runtime": 15.7439, "eval_samples_per_second": 31.758, "eval_steps_per_second": 0.254, "step": 150 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00078125, "completions/max_length": 1303.2, "completions/max_terminated_length": 494.6, "completions/mean_length": 98.003125, "completions/mean_terminated_length": 96.87923278808594, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "epoch": 0.496, "grad_norm": 0.0016110693104565144, "learning_rate": 1e-06, "loss": 0.0032, "num_tokens": 398421055.0, "reward": 0.779296875, "reward_std": 0.06850271001458168, "rewards/accuracy_reward": 0.559375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99921875, "rewards/mean_confidence_reward": 0.0, "step": 155 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 899.8, "completions/max_terminated_length": 518.4, "completions/mean_length": 93.11689453125, "completions/mean_terminated_length": 92.55267181396485, "completions/min_length": 36.0, "completions/min_terminated_length": 36.0, "epoch": 0.512, "grad_norm": 0.0013594066258519888, "learning_rate": 1e-06, "loss": 0.0022, "num_tokens": 411376556.0, "reward": 0.779443359375, "reward_std": 0.06428035944700242, "rewards/accuracy_reward": 0.55927734375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.999609375, "rewards/mean_confidence_reward": 0.0, "step": 160 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0005859375, "completions/max_length": 1116.4, "completions/max_terminated_length": 432.0, "completions/mean_length": 93.2267578125, "completions/mean_terminated_length": 92.3808364868164, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "epoch": 0.528, "grad_norm": 0.0014900276437401772, "learning_rate": 1e-06, "loss": 0.0026, "num_tokens": 424217054.0, "reward": 0.778076171875, "reward_std": 0.0658962957561016, "rewards/accuracy_reward": 0.55673828125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9994140625, "rewards/mean_confidence_reward": 0.0, "step": 165 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0005859375, "completions/max_length": 821.4, "completions/max_terminated_length": 519.6, "completions/mean_length": 93.21083984375, "completions/mean_terminated_length": 92.36523132324218, "completions/min_length": 35.2, "completions/min_terminated_length": 35.2, "epoch": 0.544, "grad_norm": 0.002942045219242573, "learning_rate": 1e-06, "loss": 0.0023, "num_tokens": 437191437.0, "reward": 0.79521484375, "reward_std": 0.07868360131978988, "rewards/accuracy_reward": 0.59111328125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99931640625, "rewards/mean_confidence_reward": 0.0, "step": 170 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0005859375, "completions/max_length": 1312.2, "completions/max_terminated_length": 406.6, "completions/mean_length": 93.5953125, "completions/mean_terminated_length": 92.75013885498046, "completions/min_length": 37.6, "completions/min_terminated_length": 37.6, "epoch": 0.56, "grad_norm": 0.0016339273424819112, "learning_rate": 1e-06, "loss": 0.0019, "num_tokens": 449827581.0, "reward": 0.768896484375, "reward_std": 0.06067224889993668, "rewards/accuracy_reward": 0.53837890625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9994140625, "rewards/mean_confidence_reward": 0.0, "step": 175 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00107421875, "completions/max_length": 1292.6, "completions/max_terminated_length": 394.6, "completions/mean_length": 94.48125, "completions/mean_terminated_length": 92.93127746582032, "completions/min_length": 43.8, "completions/min_terminated_length": 43.8, "epoch": 0.576, "grad_norm": 0.0014951933408156037, "learning_rate": 1e-06, "loss": 0.0032, "num_tokens": 462838013.0, "reward": 0.767529296875, "reward_std": 0.05619642436504364, "rewards/accuracy_reward": 0.5361328125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99892578125, "rewards/mean_confidence_reward": 0.0, "step": 180 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0001953125, "completions/max_length": 590.2, "completions/max_terminated_length": 341.8, "completions/mean_length": 90.8630859375, "completions/mean_terminated_length": 90.5811050415039, "completions/min_length": 38.6, "completions/min_terminated_length": 38.6, "epoch": 0.592, "grad_norm": 0.0016095780301839113, "learning_rate": 1e-06, "loss": 0.0011, "num_tokens": 475792483.0, "reward": 0.7609375, "reward_std": 0.05783471763134003, "rewards/accuracy_reward": 0.5220703125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9998046875, "rewards/mean_confidence_reward": 0.0, "step": 185 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0005859375, "completions/max_length": 1291.2, "completions/max_terminated_length": 429.2, "completions/mean_length": 92.3296875, "completions/mean_terminated_length": 91.4835708618164, "completions/min_length": 40.4, "completions/min_terminated_length": 40.4, "epoch": 0.608, "grad_norm": 0.0014157961122691631, "learning_rate": 1e-06, "loss": 0.0017, "num_tokens": 488593747.0, "reward": 0.775439453125, "reward_std": 0.0586208701133728, "rewards/accuracy_reward": 0.55146484375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9994140625, "rewards/mean_confidence_reward": 0.0, "step": 190 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00029296875, "completions/max_length": 938.8, "completions/max_terminated_length": 530.8, "completions/mean_length": 91.69248046875, "completions/mean_terminated_length": 91.26978149414063, "completions/min_length": 40.0, "completions/min_terminated_length": 40.0, "epoch": 0.624, "grad_norm": 0.0018766943830996752, "learning_rate": 1e-06, "loss": 0.0012, "num_tokens": 501732902.0, "reward": 0.776904296875, "reward_std": 0.05459159761667252, "rewards/accuracy_reward": 0.5541015625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99970703125, "rewards/mean_confidence_reward": 0.0, "step": 195 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00048828125, "completions/max_length": 1351.0, "completions/max_terminated_length": 483.4, "completions/mean_length": 92.82841796875, "completions/mean_terminated_length": 92.12330017089843, "completions/min_length": 41.0, "completions/min_terminated_length": 41.0, "epoch": 0.64, "grad_norm": 0.0021018637344241142, "learning_rate": 1e-06, "loss": 0.0016, "num_tokens": 514882473.0, "reward": 0.793505859375, "reward_std": 0.05022450163960457, "rewards/accuracy_reward": 0.58759765625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9994140625, "rewards/mean_confidence_reward": 0.0, "step": 200 }, { "epoch": 0.64, "eval_completions/clipped_ratio": 0.0, "eval_completions/max_length": 214.75, "eval_completions/max_terminated_length": 214.75, "eval_completions/mean_length": 91.26717376708984, "eval_completions/mean_terminated_length": 91.26717376708984, "eval_completions/min_length": 46.0, "eval_completions/min_terminated_length": 46.0, "eval_loss": 0.0, "eval_num_tokens": 514882473.0, "eval_reward": 0.7216796875, "eval_reward_std": 0.243845384567976, "eval_rewards/accuracy_reward": 0.443359375, "eval_rewards/brier_reward": 0.0, "eval_rewards/confidence_one_or_zero": 0.0, "eval_rewards/format_reward": 1.0, "eval_rewards/mean_confidence_reward": 0.0, "eval_runtime": 13.4425, "eval_samples_per_second": 37.195, "eval_steps_per_second": 0.298, "step": 200 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0001953125, "completions/max_length": 848.6, "completions/max_terminated_length": 391.0, "completions/mean_length": 91.15234375, "completions/mean_terminated_length": 90.87013397216796, "completions/min_length": 36.6, "completions/min_terminated_length": 36.6, "epoch": 0.656, "grad_norm": 0.002083825645968318, "learning_rate": 1e-06, "loss": 0.0004, "num_tokens": 527528737.0, "reward": 0.761474609375, "reward_std": 0.05676820129156113, "rewards/accuracy_reward": 0.52314453125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9998046875, "rewards/mean_confidence_reward": 0.0, "step": 205 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 1066.2, "completions/max_terminated_length": 362.2, "completions/mean_length": 93.4865234375, "completions/mean_terminated_length": 92.92289428710937, "completions/min_length": 39.0, "completions/min_terminated_length": 39.0, "epoch": 0.672, "grad_norm": 0.0021750768646597862, "learning_rate": 1e-06, "loss": 0.0016, "num_tokens": 540255799.0, "reward": 0.76650390625, "reward_std": 0.05727210938930512, "rewards/accuracy_reward": 0.5333984375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.999609375, "rewards/mean_confidence_reward": 0.0, "step": 210 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00029296875, "completions/max_length": 847.6, "completions/max_terminated_length": 399.6, "completions/mean_length": 96.2537109375, "completions/mean_terminated_length": 95.83192443847656, "completions/min_length": 39.4, "completions/min_terminated_length": 39.4, "epoch": 0.688, "grad_norm": 0.003001829609274864, "learning_rate": 1e-06, "loss": 0.0017, "num_tokens": 553051677.0, "reward": 0.781494140625, "reward_std": 0.05510343983769417, "rewards/accuracy_reward": 0.56328125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99970703125, "rewards/mean_confidence_reward": 0.0, "step": 215 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 9.765625e-05, "completions/max_length": 596.2, "completions/max_terminated_length": 408.2, "completions/mean_length": 93.3064453125, "completions/mean_terminated_length": 93.165869140625, "completions/min_length": 42.2, "completions/min_terminated_length": 42.2, "epoch": 0.704, "grad_norm": 0.0012618749169632792, "learning_rate": 1e-06, "loss": -0.0001, "num_tokens": 565729599.0, "reward": 0.782421875, "reward_std": 0.046160271018743516, "rewards/accuracy_reward": 0.56494140625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99990234375, "rewards/mean_confidence_reward": 0.0, "step": 220 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0005859375, "completions/max_length": 1132.4, "completions/max_terminated_length": 580.2, "completions/mean_length": 96.8970703125, "completions/mean_terminated_length": 96.05291748046875, "completions/min_length": 44.0, "completions/min_terminated_length": 44.0, "epoch": 0.72, "grad_norm": 0.0017941860714927316, "learning_rate": 1e-06, "loss": 0.0011, "num_tokens": 578588001.0, "reward": 0.7890625, "reward_std": 0.054417699575424194, "rewards/accuracy_reward": 0.5787109375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9994140625, "rewards/mean_confidence_reward": 0.0, "step": 225 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0005859375, "completions/max_length": 832.4, "completions/max_terminated_length": 407.2, "completions/mean_length": 97.96123046875, "completions/mean_terminated_length": 97.11712493896485, "completions/min_length": 46.6, "completions/min_terminated_length": 46.6, "epoch": 0.736, "grad_norm": 0.0014221714809536934, "learning_rate": 1e-06, "loss": 0.0012, "num_tokens": 591387028.0, "reward": 0.78837890625, "reward_std": 0.048315313458442685, "rewards/accuracy_reward": 0.57734375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9994140625, "rewards/mean_confidence_reward": 0.0, "step": 230 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0001953125, "completions/max_length": 631.4, "completions/max_terminated_length": 440.4, "completions/mean_length": 99.93115234375, "completions/mean_terminated_length": 99.65132141113281, "completions/min_length": 48.2, "completions/min_terminated_length": 48.2, "epoch": 0.752, "grad_norm": 0.0011550748022273183, "learning_rate": 1e-06, "loss": 0.0011, "num_tokens": 604493843.0, "reward": 0.785107421875, "reward_std": 0.052481997013092044, "rewards/accuracy_reward": 0.57041015625, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9998046875, "rewards/mean_confidence_reward": 0.0, "step": 235 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.001171875, "completions/max_length": 848.6, "completions/max_terminated_length": 599.0, "completions/mean_length": 105.28837890625, "completions/mean_terminated_length": 103.6119171142578, "completions/min_length": 46.2, "completions/min_terminated_length": 46.2, "epoch": 0.768, "grad_norm": 0.0019090637797489762, "learning_rate": 1e-06, "loss": 0.0023, "num_tokens": 617361020.0, "reward": 0.76201171875, "reward_std": 0.05515236109495163, "rewards/accuracy_reward": 0.5251953125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.998828125, "rewards/mean_confidence_reward": 0.0, "step": 240 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00029296875, "completions/max_length": 1095.8, "completions/max_terminated_length": 401.2, "completions/mean_length": 106.9001953125, "completions/mean_terminated_length": 106.48137512207032, "completions/min_length": 49.8, "completions/min_terminated_length": 49.8, "epoch": 0.784, "grad_norm": 0.0015024510212242603, "learning_rate": 1e-06, "loss": 0.0014, "num_tokens": 630486366.0, "reward": 0.787255859375, "reward_std": 0.05614206939935684, "rewards/accuracy_reward": 0.5748046875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99970703125, "rewards/mean_confidence_reward": 0.0, "step": 245 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00048828125, "completions/max_length": 1434.8, "completions/max_terminated_length": 533.6, "completions/mean_length": 113.03037109375, "completions/mean_terminated_length": 112.33526763916015, "completions/min_length": 51.4, "completions/min_terminated_length": 51.4, "epoch": 0.8, "grad_norm": 0.0013259610859677196, "learning_rate": 1e-06, "loss": 0.0012, "num_tokens": 643510677.0, "reward": 0.802490234375, "reward_std": 0.04771163538098335, "rewards/accuracy_reward": 0.60546875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99951171875, "rewards/mean_confidence_reward": 0.0, "step": 250 }, { "epoch": 0.8, "eval_completions/clipped_ratio": 0.0, "eval_completions/max_length": 298.75, "eval_completions/max_terminated_length": 298.75, "eval_completions/mean_length": 116.78151893615723, "eval_completions/mean_terminated_length": 116.78151893615723, "eval_completions/min_length": 58.75, "eval_completions/min_terminated_length": 58.75, "eval_loss": 0.0, "eval_num_tokens": 643510677.0, "eval_reward": 0.720703125, "eval_reward_std": 0.247483242303133, "eval_rewards/accuracy_reward": 0.44140625, "eval_rewards/brier_reward": 0.0, "eval_rewards/confidence_one_or_zero": 0.0, "eval_rewards/format_reward": 1.0, "eval_rewards/mean_confidence_reward": 0.0, "eval_runtime": 16.3536, "eval_samples_per_second": 30.574, "eval_steps_per_second": 0.245, "step": 250 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.001171875, "completions/max_length": 1216.8, "completions/max_terminated_length": 613.8, "completions/mean_length": 114.61943359375, "completions/mean_terminated_length": 112.94878845214843, "completions/min_length": 55.2, "completions/min_terminated_length": 55.2, "epoch": 0.816, "grad_norm": 0.0017014788463711739, "learning_rate": 1e-06, "loss": 0.0034, "num_tokens": 656639868.0, "reward": 0.79716796875, "reward_std": 0.059028515964746474, "rewards/accuracy_reward": 0.59560546875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99873046875, "rewards/mean_confidence_reward": 0.0, "step": 255 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00029296875, "completions/max_length": 1157.6, "completions/max_terminated_length": 482.0, "completions/mean_length": 113.28662109375, "completions/mean_terminated_length": 112.86976013183593, "completions/min_length": 50.2, "completions/min_terminated_length": 50.2, "epoch": 0.832, "grad_norm": 0.0014079577522352338, "learning_rate": 1e-06, "loss": 0.0012, "num_tokens": 669664595.0, "reward": 0.78955078125, "reward_std": 0.0593775637447834, "rewards/accuracy_reward": 0.57939453125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99970703125, "rewards/mean_confidence_reward": 0.0, "step": 260 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00048828125, "completions/max_length": 1129.8, "completions/max_terminated_length": 445.4, "completions/mean_length": 105.9974609375, "completions/mean_terminated_length": 105.29950103759765, "completions/min_length": 51.2, "completions/min_terminated_length": 51.2, "epoch": 0.848, "grad_norm": 0.002518897643312812, "learning_rate": 1e-06, "loss": 0.0019, "num_tokens": 682620697.0, "reward": 0.77509765625, "reward_std": 0.05453771948814392, "rewards/accuracy_reward": 0.55068359375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99951171875, "rewards/mean_confidence_reward": 0.0, "step": 265 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00029296875, "completions/max_length": 841.4, "completions/max_terminated_length": 388.6, "completions/mean_length": 102.1765625, "completions/mean_terminated_length": 101.75653228759765, "completions/min_length": 49.8, "completions/min_terminated_length": 49.8, "epoch": 0.864, "grad_norm": 0.002422819845378399, "learning_rate": 1e-06, "loss": 0.0015, "num_tokens": 695510121.0, "reward": 0.808837890625, "reward_std": 0.05869279354810715, "rewards/accuracy_reward": 0.61796875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99970703125, "rewards/mean_confidence_reward": 0.0, "step": 270 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00029296875, "completions/max_length": 878.2, "completions/max_terminated_length": 402.8, "completions/mean_length": 95.89345703125, "completions/mean_terminated_length": 95.47228546142578, "completions/min_length": 45.4, "completions/min_terminated_length": 45.4, "epoch": 0.88, "grad_norm": 0.002903913613408804, "learning_rate": 1e-06, "loss": 0.0017, "num_tokens": 708495462.0, "reward": 0.7587890625, "reward_std": 0.05212245061993599, "rewards/accuracy_reward": 0.51796875, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.999609375, "rewards/mean_confidence_reward": 0.0, "step": 275 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00048828125, "completions/max_length": 1318.2, "completions/max_terminated_length": 486.4, "completions/mean_length": 93.419921875, "completions/mean_terminated_length": 92.71511383056641, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "epoch": 0.896, "grad_norm": 0.001861903234384954, "learning_rate": 1e-06, "loss": 0.0021, "num_tokens": 721419250.0, "reward": 0.777197265625, "reward_std": 0.046849222481250764, "rewards/accuracy_reward": 0.5548828125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99951171875, "rewards/mean_confidence_reward": 0.0, "step": 280 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00068359375, "completions/max_length": 1306.6, "completions/max_terminated_length": 360.6, "completions/mean_length": 90.81923828125, "completions/mean_terminated_length": 89.83069915771485, "completions/min_length": 43.2, "completions/min_terminated_length": 43.2, "epoch": 0.912, "grad_norm": 0.0009662771481089294, "learning_rate": 1e-06, "loss": 0.0016, "num_tokens": 734256855.0, "reward": 0.782080078125, "reward_std": 0.04545313939452171, "rewards/accuracy_reward": 0.56484375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99931640625, "rewards/mean_confidence_reward": 0.0, "step": 285 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00029296875, "completions/max_length": 816.4, "completions/max_terminated_length": 363.8, "completions/mean_length": 86.1810546875, "completions/mean_terminated_length": 85.75641479492188, "completions/min_length": 37.0, "completions/min_terminated_length": 37.0, "epoch": 0.928, "grad_norm": 0.0012876720866188407, "learning_rate": 1e-06, "loss": 0.0012, "num_tokens": 747022485.0, "reward": 0.772705078125, "reward_std": 0.045815450698137285, "rewards/accuracy_reward": 0.545703125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99970703125, "rewards/mean_confidence_reward": 0.0, "step": 290 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.0001953125, "completions/max_length": 822.2, "completions/max_terminated_length": 324.0, "completions/mean_length": 83.45751953125, "completions/mean_terminated_length": 83.17401275634765, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "epoch": 0.944, "grad_norm": 0.002664135303348303, "learning_rate": 1e-06, "loss": 0.0011, "num_tokens": 759708834.0, "reward": 0.779052734375, "reward_std": 0.05606053844094276, "rewards/accuracy_reward": 0.55830078125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.9998046875, "rewards/mean_confidence_reward": 0.0, "step": 295 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 1058.4, "completions/max_terminated_length": 550.4, "completions/mean_length": 83.72802734375, "completions/mean_terminated_length": 83.16080169677734, "completions/min_length": 41.0, "completions/min_terminated_length": 41.0, "epoch": 0.96, "grad_norm": 0.0016519302735105157, "learning_rate": 1e-06, "loss": 0.0008, "num_tokens": 772362849.0, "reward": 0.7677734375, "reward_std": 0.04194698035717011, "rewards/accuracy_reward": 0.5359375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.999609375, "rewards/mean_confidence_reward": 0.0, "step": 300 }, { "epoch": 0.96, "eval_completions/clipped_ratio": 0.0, "eval_completions/max_length": 254.0, "eval_completions/max_terminated_length": 254.0, "eval_completions/mean_length": 84.06842613220215, "eval_completions/mean_terminated_length": 84.06842613220215, "eval_completions/min_length": 45.25, "eval_completions/min_terminated_length": 45.25, "eval_loss": 0.0, "eval_num_tokens": 772362849.0, "eval_reward": 0.72265625, "eval_reward_std": 0.2485281154513359, "eval_rewards/accuracy_reward": 0.4453125, "eval_rewards/brier_reward": 0.0, "eval_rewards/confidence_one_or_zero": 0.0, "eval_rewards/format_reward": 1.0, "eval_rewards/mean_confidence_reward": 0.0, "eval_runtime": 14.5394, "eval_samples_per_second": 34.389, "eval_steps_per_second": 0.275, "step": 300 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000390625, "completions/max_length": 1089.0, "completions/max_terminated_length": 383.0, "completions/mean_length": 83.73583984375, "completions/mean_terminated_length": 83.16815948486328, "completions/min_length": 41.6, "completions/min_terminated_length": 41.6, "epoch": 0.976, "grad_norm": 0.001557100797072053, "learning_rate": 1e-06, "loss": 0.0009, "num_tokens": 784937744.0, "reward": 0.784521484375, "reward_std": 0.051283557713031766, "rewards/accuracy_reward": 0.56943359375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.999609375, "rewards/mean_confidence_reward": 0.0, "step": 305 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.00029296875, "completions/max_length": 807.2, "completions/max_terminated_length": 342.4, "completions/mean_length": 85.384375, "completions/mean_terminated_length": 84.96044158935547, "completions/min_length": 40.6, "completions/min_terminated_length": 40.6, "epoch": 0.992, "grad_norm": 0.0014634578255936503, "learning_rate": 1e-06, "loss": 0.0006, "num_tokens": 797796880.0, "reward": 0.776220703125, "reward_std": 0.04600930213928223, "rewards/accuracy_reward": 0.552734375, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.99970703125, "rewards/mean_confidence_reward": 0.0, "step": 310 }, { "clip_ratio/high_max": 0.0, "clip_ratio/high_mean": 0.0, "clip_ratio/low_mean": 0.0, "clip_ratio/low_min": 0.0, "clip_ratio/region_mean": 0.0, "completions/clipped_ratio": 0.000244140625, "completions/max_length": 943.0, "completions/max_terminated_length": 318.5, "completions/mean_length": 85.54854202270508, "completions/mean_terminated_length": 85.19485473632812, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "epoch": 0.9984, "num_tokens": 802903205.0, "reward": 0.7908935546875, "reward_std": 0.050602177157998085, "rewards/accuracy_reward": 0.58251953125, "rewards/brier_reward": 0.0, "rewards/confidence_one_or_zero": 0.0, "rewards/format_reward": 0.999267578125, "rewards/mean_confidence_reward": 0.0, "step": 312, "total_flos": 0.0, "train_loss": 0.003547575732227415, "train_runtime": 65955.8089, "train_samples_per_second": 0.303, "train_steps_per_second": 0.005 } ], "logging_steps": 5, "max_steps": 312, "num_input_tokens_seen": 802903205, "num_train_epochs": 1, "save_steps": 60, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": true }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 8, "trial_name": null, "trial_params": null }