{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.08, "eval_steps": 500, "global_step": 800, "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": 0.0, "completions/max_length": 382.0, "completions/max_terminated_length": 382.0, "completions/mean_length": 287.875, "completions/mean_terminated_length": 287.875, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "entropy": 0.1642230749130249, "epoch": 0.0001, "frac_reward_zero_std": 0.0, "grad_norm": 0.665309488773346, "learning_rate": 1e-05, "loss": -0.11556684970855713, "num_tokens": 3951.0, "reward": 0.125, "reward_std": 0.3535533845424652, "rewards/boxed_accuracy_reward/mean": 0.125, "rewards/boxed_accuracy_reward/std": 0.3535533845424652, "step": 1, "step_time": 26.083515274000092 }, { "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": 438.1111111111111, "completions/max_terminated_length": 438.1111111111111, "completions/mean_length": 253.38888888888889, "completions/mean_terminated_length": 253.38888888888889, "completions/min_length": 132.33333333333334, "completions/min_terminated_length": 132.33333333333334, "entropy": 0.19083112478256226, "epoch": 0.001, "frac_reward_zero_std": 0.3333333333333333, "grad_norm": 0.7672234177589417, "learning_rate": 9.91e-06, "loss": 0.017603478497929044, "num_tokens": 34851.0, "reward": 0.1527777777777778, "reward_std": 0.2903762327300178, "rewards/boxed_accuracy_reward/mean": 0.1527777777777778, "rewards/boxed_accuracy_reward/std": 0.2903762393527561, "step": 10, "step_time": 29.407547047666917 }, { "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": 0.0625, "completions/max_length": 598.4, "completions/max_terminated_length": 454.2, "completions/mean_length": 326.7125, "completions/mean_terminated_length": 281.9235778808594, "completions/min_length": 157.3, "completions/min_terminated_length": 157.3, "entropy": 0.22966246232390403, "epoch": 0.002, "frac_reward_zero_std": 0.7, "grad_norm": 0.0, "learning_rate": 9.810000000000001e-06, "loss": 0.015656116604804992, "num_tokens": 74804.0, "reward": 0.15, "reward_std": 0.12416292428970337, "rewards/boxed_accuracy_reward/mean": 0.15, "rewards/boxed_accuracy_reward/std": 0.12416292428970337, "step": 20, "step_time": 40.079171687399864 }, { "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": 0.0125, "completions/max_length": 471.4, "completions/max_terminated_length": 418.5, "completions/mean_length": 261.45, "completions/mean_terminated_length": 251.90892944335937, "completions/min_length": 161.7, "completions/min_terminated_length": 161.7, "entropy": 0.17057178877294063, "epoch": 0.003, "frac_reward_zero_std": 0.8, "grad_norm": 0.0, "learning_rate": 9.71e-06, "loss": -0.01473381370306015, "num_tokens": 109952.0, "reward": 0.25, "reward_std": 0.08711025416851044, "rewards/boxed_accuracy_reward/mean": 0.25, "rewards/boxed_accuracy_reward/std": 0.08711026012897491, "step": 30, "step_time": 31.344073937199575 }, { "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": 0.0375, "completions/max_length": 462.8, "completions/max_terminated_length": 416.3, "completions/mean_length": 292.6375, "completions/mean_terminated_length": 267.9024993896484, "completions/min_length": 172.9, "completions/min_terminated_length": 172.9, "entropy": 0.22824809104204177, "epoch": 0.004, "frac_reward_zero_std": 0.8, "grad_norm": 0.0, "learning_rate": 9.610000000000001e-06, "loss": 0.014347019791603088, "num_tokens": 148715.0, "reward": 0.1, "reward_std": 0.09258201122283935, "rewards/boxed_accuracy_reward/mean": 0.1, "rewards/boxed_accuracy_reward/std": 0.09258201122283935, "step": 40, "step_time": 30.823996917499972 }, { "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": 0.025, "completions/max_length": 543.5, "completions/max_terminated_length": 415.6, "completions/mean_length": 269.2375, "completions/mean_terminated_length": 249.4482162475586, "completions/min_length": 160.6, "completions/min_terminated_length": 160.6, "entropy": 0.266183939576149, "epoch": 0.005, "frac_reward_zero_std": 0.4, "grad_norm": 0.7730127573013306, "learning_rate": 9.51e-06, "loss": 0.047160887718200685, "num_tokens": 184662.0, "reward": 0.1625, "reward_std": 0.26849985122680664, "rewards/boxed_accuracy_reward/mean": 0.1625, "rewards/boxed_accuracy_reward/std": 0.2684998571872711, "step": 50, "step_time": 36.26071728879979 }, { "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": 445.4, "completions/max_terminated_length": 445.4, "completions/mean_length": 271.05, "completions/mean_terminated_length": 271.05, "completions/min_length": 168.9, "completions/min_terminated_length": 168.9, "entropy": 0.18989864736795425, "epoch": 0.006, "frac_reward_zero_std": 0.5, "grad_norm": 0.0, "learning_rate": 9.41e-06, "loss": 0.07637788653373719, "num_tokens": 221874.0, "reward": 0.25, "reward_std": 0.2276727557182312, "rewards/boxed_accuracy_reward/mean": 0.25, "rewards/boxed_accuracy_reward/std": 0.22767276763916017, "step": 60, "step_time": 29.8493166889999 }, { "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": 0.025, "completions/max_length": 559.9, "completions/max_terminated_length": 456.6, "completions/mean_length": 279.3125, "completions/mean_terminated_length": 262.7035736083984, "completions/min_length": 157.0, "completions/min_terminated_length": 157.0, "entropy": 0.2535190790891647, "epoch": 0.007, "frac_reward_zero_std": 0.9, "grad_norm": 0.0, "learning_rate": 9.31e-06, "loss": 0.007473611086606979, "num_tokens": 257171.0, "reward": 0.05, "reward_std": 0.05345224738121033, "rewards/boxed_accuracy_reward/mean": 0.05, "rewards/boxed_accuracy_reward/std": 0.05345224738121033, "step": 70, "step_time": 37.422609474000275 }, { "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": 0.025, "completions/max_length": 506.5, "completions/max_terminated_length": 375.0, "completions/mean_length": 246.025, "completions/mean_terminated_length": 226.50714569091798, "completions/min_length": 146.6, "completions/min_terminated_length": 146.6, "entropy": 0.21952729374170304, "epoch": 0.008, "frac_reward_zero_std": 0.6, "grad_norm": 0.0, "learning_rate": 9.210000000000002e-06, "loss": -0.004495994746685028, "num_tokens": 291301.0, "reward": 0.175, "reward_std": 0.18685350716114044, "rewards/boxed_accuracy_reward/mean": 0.175, "rewards/boxed_accuracy_reward/std": 0.18685351312160492, "step": 80, "step_time": 33.85629228980069 }, { "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": 417.1, "completions/max_terminated_length": 417.1, "completions/mean_length": 270.3375, "completions/mean_terminated_length": 270.3375, "completions/min_length": 177.8, "completions/min_terminated_length": 177.8, "entropy": 0.2391076095402241, "epoch": 0.009, "frac_reward_zero_std": 0.5, "grad_norm": 0.6646445393562317, "learning_rate": 9.110000000000001e-06, "loss": 0.0037924151867628096, "num_tokens": 326552.0, "reward": 0.1125, "reward_std": 0.2095758467912674, "rewards/boxed_accuracy_reward/mean": 0.1125, "rewards/boxed_accuracy_reward/std": 0.20957585871219636, "step": 90, "step_time": 27.908505431900267 }, { "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": 0.0625, "completions/max_length": 489.8, "completions/max_terminated_length": 405.3, "completions/mean_length": 305.5, "completions/mean_terminated_length": 269.0066680908203, "completions/min_length": 142.9, "completions/min_terminated_length": 142.9, "entropy": 0.2832915961742401, "epoch": 0.01, "frac_reward_zero_std": 0.6, "grad_norm": 0.0, "learning_rate": 9.01e-06, "loss": 0.01493358165025711, "num_tokens": 364440.0, "reward": 0.2, "reward_std": 0.17045392990112304, "rewards/boxed_accuracy_reward/mean": 0.2, "rewards/boxed_accuracy_reward/std": 0.17045392990112304, "step": 100, "step_time": 32.493894719798845 }, { "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": 0.0125, "completions/max_length": 454.8, "completions/max_terminated_length": 377.0, "completions/mean_length": 258.625, "completions/mean_terminated_length": 248.3375015258789, "completions/min_length": 170.1, "completions/min_terminated_length": 170.1, "entropy": 0.2617956634610891, "epoch": 0.011, "frac_reward_zero_std": 0.5, "grad_norm": 0.29700157046318054, "learning_rate": 8.910000000000001e-06, "loss": 0.03313029408454895, "num_tokens": 398914.0, "reward": 0.125, "reward_std": 0.22598327100276946, "rewards/boxed_accuracy_reward/mean": 0.125, "rewards/boxed_accuracy_reward/std": 0.22598327696323395, "step": 110, "step_time": 30.349950915500084 }, { "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": 0.0125, "completions/max_length": 457.5, "completions/max_terminated_length": 389.1, "completions/mean_length": 265.4875, "completions/mean_terminated_length": 255.85714416503907, "completions/min_length": 164.2, "completions/min_terminated_length": 164.2, "entropy": 0.2188379220664501, "epoch": 0.012, "frac_reward_zero_std": 0.8, "grad_norm": 0.0, "learning_rate": 8.81e-06, "loss": 0.02148536443710327, "num_tokens": 434921.0, "reward": 0.0625, "reward_std": 0.08880758583545685, "rewards/boxed_accuracy_reward/mean": 0.0625, "rewards/boxed_accuracy_reward/std": 0.08880758583545685, "step": 120, "step_time": 30.65767741319978 }, { "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": 0.0125, "completions/max_length": 424.0, "completions/max_terminated_length": 382.3, "completions/mean_length": 237.0125, "completions/mean_terminated_length": 228.27142944335938, "completions/min_length": 139.3, "completions/min_terminated_length": 139.3, "entropy": 0.2461308181285858, "epoch": 0.013, "frac_reward_zero_std": 0.5, "grad_norm": 0.0, "learning_rate": 8.710000000000001e-06, "loss": 0.013855263590812683, "num_tokens": 468658.0, "reward": 0.1, "reward_std": 0.20411193668842315, "rewards/boxed_accuracy_reward/mean": 0.1, "rewards/boxed_accuracy_reward/std": 0.20411194264888763, "step": 130, "step_time": 28.27658463789994 }, { "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": 0.0125, "completions/max_length": 416.1, "completions/max_terminated_length": 357.2, "completions/mean_length": 233.2625, "completions/mean_terminated_length": 223.3732147216797, "completions/min_length": 130.1, "completions/min_terminated_length": 130.1, "entropy": 0.29308837205171584, "epoch": 0.014, "frac_reward_zero_std": 0.5, "grad_norm": 0.0, "learning_rate": 8.61e-06, "loss": 0.0346070647239685, "num_tokens": 500159.0, "reward": 0.1875, "reward_std": 0.22768060266971588, "rewards/boxed_accuracy_reward/mean": 0.1875, "rewards/boxed_accuracy_reward/std": 0.22768060266971588, "step": 140, "step_time": 27.783400436999727 }, { "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": 503.8, "completions/max_terminated_length": 503.8, "completions/mean_length": 295.05, "completions/mean_terminated_length": 295.05, "completions/min_length": 184.6, "completions/min_terminated_length": 184.6, "entropy": 0.20557359904050826, "epoch": 0.015, "frac_reward_zero_std": 0.7, "grad_norm": 0.0, "learning_rate": 8.51e-06, "loss": 0.03165164291858673, "num_tokens": 537435.0, "reward": 0.1625, "reward_std": 0.14433692693710326, "rewards/boxed_accuracy_reward/mean": 0.1625, "rewards/boxed_accuracy_reward/std": 0.14433693289756774, "step": 150, "step_time": 33.749151340000026 }, { "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": 0.025, "completions/max_length": 504.0, "completions/max_terminated_length": 405.2, "completions/mean_length": 280.05, "completions/mean_terminated_length": 261.85714569091795, "completions/min_length": 158.9, "completions/min_terminated_length": 158.9, "entropy": 0.2418084606528282, "epoch": 0.016, "frac_reward_zero_std": 0.4, "grad_norm": 0.9073280096054077, "learning_rate": 8.41e-06, "loss": 0.029856789112091064, "num_tokens": 573351.0, "reward": 0.2, "reward_std": 0.2794355183839798, "rewards/boxed_accuracy_reward/mean": 0.2, "rewards/boxed_accuracy_reward/std": 0.2794355243444443, "step": 160, "step_time": 36.18624399920045 }, { "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": 484.3, "completions/max_terminated_length": 484.3, "completions/mean_length": 281.0, "completions/mean_terminated_length": 281.0, "completions/min_length": 167.8, "completions/min_terminated_length": 167.8, "entropy": 0.21754518747329712, "epoch": 0.017, "frac_reward_zero_std": 0.9, "grad_norm": 0.0, "learning_rate": 8.31e-06, "loss": 0.01904902458190918, "num_tokens": 610279.0, "reward": 0.125, "reward_std": 0.046291005611419675, "rewards/boxed_accuracy_reward/mean": 0.125, "rewards/boxed_accuracy_reward/std": 0.046291005611419675, "step": 170, "step_time": 35.24356591649958 }, { "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": 0.0125, "completions/max_length": 472.6, "completions/max_terminated_length": 421.2, "completions/mean_length": 269.1625, "completions/mean_terminated_length": 260.52142944335935, "completions/min_length": 159.9, "completions/min_terminated_length": 159.9, "entropy": 0.22995990961790086, "epoch": 0.018, "frac_reward_zero_std": 0.8, "grad_norm": 0.0, "learning_rate": 8.210000000000001e-06, "loss": -0.011320264637470245, "num_tokens": 646660.0, "reward": 0.0625, "reward_std": 0.08880758583545685, "rewards/boxed_accuracy_reward/mean": 0.0625, "rewards/boxed_accuracy_reward/std": 0.08880758583545685, "step": 180, "step_time": 35.16620119949948 }, { "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": 0.0375, "completions/max_length": 570.6, "completions/max_terminated_length": 513.5, "completions/mean_length": 297.2625, "completions/mean_terminated_length": 272.977978515625, "completions/min_length": 156.5, "completions/min_terminated_length": 156.5, "entropy": 0.24057478606700897, "epoch": 0.019, "frac_reward_zero_std": 0.5, "grad_norm": 1.0508952140808105, "learning_rate": 8.110000000000002e-06, "loss": 0.01611926555633545, "num_tokens": 685673.0, "reward": 0.1, "reward_std": 0.1948736011981964, "rewards/boxed_accuracy_reward/mean": 0.1, "rewards/boxed_accuracy_reward/std": 0.1948736011981964, "step": 190, "step_time": 39.78827415060077 }, { "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": 0.0125, "completions/max_length": 379.3, "completions/max_terminated_length": 322.7, "completions/mean_length": 236.0875, "completions/mean_terminated_length": 227.3767883300781, "completions/min_length": 150.7, "completions/min_terminated_length": 150.7, "entropy": 0.21252653896808624, "epoch": 0.02, "frac_reward_zero_std": 0.4, "grad_norm": 0.7992851138114929, "learning_rate": 8.010000000000001e-06, "loss": 0.02353036105632782, "num_tokens": 719336.0, "reward": 0.35, "reward_std": 0.30299633741378784, "rewards/boxed_accuracy_reward/mean": 0.35, "rewards/boxed_accuracy_reward/std": 0.3029963493347168, "step": 200, "step_time": 27.117062961000194 }, { "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": 0.0875, "completions/max_length": 609.6, "completions/max_terminated_length": 431.3, "completions/mean_length": 345.1875, "completions/mean_terminated_length": 294.5657165527344, "completions/min_length": 165.6, "completions/min_terminated_length": 165.6, "entropy": 0.17211427241563798, "epoch": 0.021, "frac_reward_zero_std": 0.9, "grad_norm": 0.0, "learning_rate": 7.91e-06, "loss": -0.0028946617618203162, "num_tokens": 759991.0, "reward": 0.0125, "reward_std": 0.03535533845424652, "rewards/boxed_accuracy_reward/mean": 0.0125, "rewards/boxed_accuracy_reward/std": 0.03535533845424652, "step": 210, "step_time": 44.109834954999315 }, { "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": 0.025, "completions/max_length": 562.9, "completions/max_terminated_length": 454.9, "completions/mean_length": 275.2125, "completions/mean_terminated_length": 257.24286193847655, "completions/min_length": 149.9, "completions/min_terminated_length": 149.9, "entropy": 0.2036581441760063, "epoch": 0.022, "frac_reward_zero_std": 0.6, "grad_norm": 0.9508808255195618, "learning_rate": 7.810000000000001e-06, "loss": 0.05190147757530213, "num_tokens": 795232.0, "reward": 0.2375, "reward_std": 0.17422835528850555, "rewards/boxed_accuracy_reward/mean": 0.2375, "rewards/boxed_accuracy_reward/std": 0.17422835528850555, "step": 220, "step_time": 38.12656000370043 }, { "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": 0.0125, "completions/max_length": 372.2, "completions/max_terminated_length": 322.1, "completions/mean_length": 216.725, "completions/mean_terminated_length": 206.85714416503907, "completions/min_length": 123.2, "completions/min_terminated_length": 123.2, "entropy": 0.24748676642775536, "epoch": 0.023, "frac_reward_zero_std": 0.6, "grad_norm": 0.825386643409729, "learning_rate": 7.71e-06, "loss": -0.019961801171302796, "num_tokens": 828634.0, "reward": 0.225, "reward_std": 0.1632926881313324, "rewards/boxed_accuracy_reward/mean": 0.225, "rewards/boxed_accuracy_reward/std": 0.1632926881313324, "step": 230, "step_time": 24.96606878720049 }, { "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": 373.3, "completions/max_terminated_length": 373.3, "completions/mean_length": 232.2, "completions/mean_terminated_length": 232.2, "completions/min_length": 150.1, "completions/min_terminated_length": 150.1, "entropy": 0.2603755287826061, "epoch": 0.024, "frac_reward_zero_std": 0.4, "grad_norm": 0.5379490256309509, "learning_rate": 7.610000000000001e-06, "loss": 0.03033584952354431, "num_tokens": 861162.0, "reward": 0.1875, "reward_std": 0.2722664296627045, "rewards/boxed_accuracy_reward/mean": 0.1875, "rewards/boxed_accuracy_reward/std": 0.2722664475440979, "step": 240, "step_time": 24.956236593399808 }, { "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": 0.0125, "completions/max_length": 532.6, "completions/max_terminated_length": 465.1, "completions/mean_length": 262.9375, "completions/mean_terminated_length": 252.74464416503906, "completions/min_length": 135.1, "completions/min_terminated_length": 135.1, "entropy": 0.2639094963669777, "epoch": 0.025, "frac_reward_zero_std": 0.8, "grad_norm": 0.0, "learning_rate": 7.510000000000001e-06, "loss": 0.014231480658054352, "num_tokens": 896165.0, "reward": 0.075, "reward_std": 0.08711025416851044, "rewards/boxed_accuracy_reward/mean": 0.075, "rewards/boxed_accuracy_reward/std": 0.08711026012897491, "step": 250, "step_time": 35.73487980219979 }, { "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": 416.4, "completions/max_terminated_length": 416.4, "completions/mean_length": 263.975, "completions/mean_terminated_length": 263.975, "completions/min_length": 163.7, "completions/min_terminated_length": 163.7, "entropy": 0.249638444930315, "epoch": 0.026, "frac_reward_zero_std": 0.6, "grad_norm": 0.0, "learning_rate": 7.41e-06, "loss": 0.01857263296842575, "num_tokens": 930771.0, "reward": 0.1625, "reward_std": 0.15235702097415924, "rewards/boxed_accuracy_reward/mean": 0.1625, "rewards/boxed_accuracy_reward/std": 0.15235702097415924, "step": 260, "step_time": 28.139153176400214 }, { "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": 422.7, "completions/max_terminated_length": 422.7, "completions/mean_length": 251.9875, "completions/mean_terminated_length": 251.9875, "completions/min_length": 153.3, "completions/min_terminated_length": 153.3, "entropy": 0.24129935801029206, "epoch": 0.027, "frac_reward_zero_std": 0.7, "grad_norm": 0.0, "learning_rate": 7.31e-06, "loss": -0.032139381766319274, "num_tokens": 963458.0, "reward": 0.2125, "reward_std": 0.14433692693710326, "rewards/boxed_accuracy_reward/mean": 0.2125, "rewards/boxed_accuracy_reward/std": 0.14433693289756774, "step": 270, "step_time": 28.544226764599806 }, { "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": 332.8, "completions/max_terminated_length": 332.8, "completions/mean_length": 231.9125, "completions/mean_terminated_length": 231.9125, "completions/min_length": 148.2, "completions/min_terminated_length": 148.2, "entropy": 0.26727280877530574, "epoch": 0.028, "frac_reward_zero_std": 0.6, "grad_norm": 0.9024311900138855, "learning_rate": 7.2100000000000004e-06, "loss": 0.021772387623786926, "num_tokens": 996579.0, "reward": 0.15, "reward_std": 0.17422050833702088, "rewards/boxed_accuracy_reward/mean": 0.15, "rewards/boxed_accuracy_reward/std": 0.17422052025794982, "step": 280, "step_time": 22.47074644069944 }, { "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": 0.0625, "completions/max_length": 443.4, "completions/max_terminated_length": 364.0, "completions/mean_length": 265.1375, "completions/mean_terminated_length": 211.40833435058593, "completions/min_length": 125.6, "completions/min_terminated_length": 125.6, "entropy": 0.2202783677726984, "epoch": 0.029, "frac_reward_zero_std": 0.5, "grad_norm": 0.23354178667068481, "learning_rate": 7.1100000000000005e-06, "loss": 0.039704719185829164, "num_tokens": 1033446.0, "reward": 0.2375, "reward_std": 0.2478467583656311, "rewards/boxed_accuracy_reward/mean": 0.2375, "rewards/boxed_accuracy_reward/std": 0.24784677624702453, "step": 290, "step_time": 29.699856723000266 }, { "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": 403.5, "completions/max_terminated_length": 403.5, "completions/mean_length": 281.5125, "completions/mean_terminated_length": 281.5125, "completions/min_length": 162.4, "completions/min_terminated_length": 162.4, "entropy": 0.2921067249029875, "epoch": 0.03, "frac_reward_zero_std": 0.5, "grad_norm": 0.0, "learning_rate": 7.01e-06, "loss": 0.006478501856327057, "num_tokens": 1071503.0, "reward": 0.2625, "reward_std": 0.19864802658557892, "rewards/boxed_accuracy_reward/mean": 0.2625, "rewards/boxed_accuracy_reward/std": 0.19864802658557892, "step": 300, "step_time": 27.02418423309937 }, { "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": 438.9, "completions/max_terminated_length": 438.9, "completions/mean_length": 247.8, "completions/mean_terminated_length": 247.8, "completions/min_length": 139.4, "completions/min_terminated_length": 139.4, "entropy": 0.22176421582698821, "epoch": 0.031, "frac_reward_zero_std": 0.4, "grad_norm": 1.1117596626281738, "learning_rate": 6.91e-06, "loss": -0.027645233273506164, "num_tokens": 1109583.0, "reward": 0.1625, "reward_std": 0.26302809417247774, "rewards/boxed_accuracy_reward/mean": 0.1625, "rewards/boxed_accuracy_reward/std": 0.26302810609340666, "step": 310, "step_time": 29.716965569899912 }, { "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": 0.0125, "completions/max_length": 388.2, "completions/max_terminated_length": 343.9, "completions/mean_length": 237.475, "completions/mean_terminated_length": 228.22321472167968, "completions/min_length": 132.2, "completions/min_terminated_length": 132.2, "entropy": 0.249158176779747, "epoch": 0.032, "frac_reward_zero_std": 0.7, "grad_norm": 0.0, "learning_rate": 6.810000000000001e-06, "loss": -0.009709689766168594, "num_tokens": 1142949.0, "reward": 0.2, "reward_std": 0.11700168251991272, "rewards/boxed_accuracy_reward/mean": 0.2, "rewards/boxed_accuracy_reward/std": 0.11700168251991272, "step": 320, "step_time": 26.217313883200042 }, { "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": 355.9, "completions/max_terminated_length": 355.9, "completions/mean_length": 207.9625, "completions/mean_terminated_length": 207.9625, "completions/min_length": 134.0, "completions/min_terminated_length": 134.0, "entropy": 0.257133349776268, "epoch": 0.033, "frac_reward_zero_std": 0.6, "grad_norm": 0.0, "learning_rate": 6.710000000000001e-06, "loss": 0.0044826693832874295, "num_tokens": 1177898.0, "reward": 0.325, "reward_std": 0.17422050833702088, "rewards/boxed_accuracy_reward/mean": 0.325, "rewards/boxed_accuracy_reward/std": 0.17422052025794982, "step": 330, "step_time": 24.09444000950098 }, { "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": 0.0125, "completions/max_length": 455.1, "completions/max_terminated_length": 405.7, "completions/mean_length": 259.3375, "completions/mean_terminated_length": 250.85535888671876, "completions/min_length": 135.2, "completions/min_terminated_length": 135.2, "entropy": 0.2380140107125044, "epoch": 0.034, "frac_reward_zero_std": 0.4, "grad_norm": 0.0, "learning_rate": 6.610000000000001e-06, "loss": 0.07607665061950683, "num_tokens": 1214925.0, "reward": 0.375, "reward_std": 0.25040294229984283, "rewards/boxed_accuracy_reward/mean": 0.375, "rewards/boxed_accuracy_reward/std": 0.2504029482603073, "step": 340, "step_time": 30.860040666599524 }, { "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": 0.0375, "completions/max_length": 509.9, "completions/max_terminated_length": 390.4, "completions/mean_length": 283.525, "completions/mean_terminated_length": 255.67381134033204, "completions/min_length": 167.1, "completions/min_terminated_length": 167.1, "entropy": 0.2493958815932274, "epoch": 0.035, "frac_reward_zero_std": 0.7, "grad_norm": 0.0, "learning_rate": 6.51e-06, "loss": 0.040489372611045835, "num_tokens": 1249775.0, "reward": 0.1125, "reward_std": 0.15526474714279176, "rewards/boxed_accuracy_reward/mean": 0.1125, "rewards/boxed_accuracy_reward/std": 0.15526476502418518, "step": 350, "step_time": 34.709320317299714 }, { "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": 379.4, "completions/max_terminated_length": 379.4, "completions/mean_length": 231.2125, "completions/mean_terminated_length": 231.2125, "completions/min_length": 124.0, "completions/min_terminated_length": 124.0, "entropy": 0.268961301445961, "epoch": 0.036, "frac_reward_zero_std": 0.8, "grad_norm": 0.0, "learning_rate": 6.4100000000000005e-06, "loss": -0.025786453485488893, "num_tokens": 1284312.0, "reward": 0.1, "reward_std": 0.07071067690849304, "rewards/boxed_accuracy_reward/mean": 0.1, "rewards/boxed_accuracy_reward/std": 0.07071067690849304, "step": 360, "step_time": 25.807292600200164 }, { "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": 331.3, "completions/max_terminated_length": 331.3, "completions/mean_length": 220.925, "completions/mean_terminated_length": 220.925, "completions/min_length": 154.9, "completions/min_terminated_length": 154.9, "entropy": 0.20115781389176846, "epoch": 0.037, "frac_reward_zero_std": 0.7, "grad_norm": 0.40251433849334717, "learning_rate": 6.3100000000000006e-06, "loss": -0.008264368772506714, "num_tokens": 1314754.0, "reward": 0.1125, "reward_std": 0.10606601536273956, "rewards/boxed_accuracy_reward/mean": 0.1125, "rewards/boxed_accuracy_reward/std": 0.10606601536273956, "step": 370, "step_time": 22.319113775398726 }, { "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": 0.0125, "completions/max_length": 397.9, "completions/max_terminated_length": 378.2, "completions/mean_length": 229.525, "completions/mean_terminated_length": 221.51964416503907, "completions/min_length": 131.6, "completions/min_terminated_length": 131.6, "entropy": 0.23055790960788727, "epoch": 0.038, "frac_reward_zero_std": 0.9, "grad_norm": 0.0, "learning_rate": 6.210000000000001e-06, "loss": 0.020930925011634828, "num_tokens": 1347092.0, "reward": 0.025, "reward_std": 0.046291005611419675, "rewards/boxed_accuracy_reward/mean": 0.025, "rewards/boxed_accuracy_reward/std": 0.046291005611419675, "step": 380, "step_time": 26.844786134899188 }, { "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": 0.0125, "completions/max_length": 486.9, "completions/max_terminated_length": 446.0, "completions/mean_length": 300.1, "completions/mean_terminated_length": 291.9678588867188, "completions/min_length": 183.8, "completions/min_terminated_length": 183.8, "entropy": 0.1606756892055273, "epoch": 0.039, "frac_reward_zero_std": 0.5, "grad_norm": 0.0, "learning_rate": 6.110000000000001e-06, "loss": -0.025502890348434448, "num_tokens": 1385580.0, "reward": 0.1875, "reward_std": 0.245769664645195, "rewards/boxed_accuracy_reward/mean": 0.1875, "rewards/boxed_accuracy_reward/std": 0.24576967656612397, "step": 390, "step_time": 32.540994760800096 }, { "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": 311.1, "completions/max_terminated_length": 311.1, "completions/mean_length": 216.7125, "completions/mean_terminated_length": 216.7125, "completions/min_length": 156.9, "completions/min_terminated_length": 156.9, "entropy": 0.19945215657353402, "epoch": 0.04, "frac_reward_zero_std": 0.6, "grad_norm": 0.6455156803131104, "learning_rate": 6.01e-06, "loss": 0.026773136854171754, "num_tokens": 1417461.0, "reward": 0.225, "reward_std": 0.1632926881313324, "rewards/boxed_accuracy_reward/mean": 0.225, "rewards/boxed_accuracy_reward/std": 0.1632926881313324, "step": 400, "step_time": 21.00104573589997 }, { "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": 0.05, "completions/max_length": 493.7, "completions/max_terminated_length": 403.2, "completions/mean_length": 278.675, "completions/mean_terminated_length": 238.26178894042968, "completions/min_length": 146.1, "completions/min_terminated_length": 146.1, "entropy": 0.20032316148281099, "epoch": 0.041, "frac_reward_zero_std": 0.6, "grad_norm": 0.0, "learning_rate": 5.91e-06, "loss": 0.045339572429656985, "num_tokens": 1456083.0, "reward": 0.1375, "reward_std": 0.15235702097415924, "rewards/boxed_accuracy_reward/mean": 0.1375, "rewards/boxed_accuracy_reward/std": 0.15235702097415924, "step": 410, "step_time": 33.424314678498924 }, { "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": 0.0125, "completions/max_length": 380.4, "completions/max_terminated_length": 327.8, "completions/mean_length": 234.2375, "completions/mean_terminated_length": 225.44107360839843, "completions/min_length": 155.1, "completions/min_terminated_length": 155.1, "entropy": 0.20021430365741252, "epoch": 0.042, "frac_reward_zero_std": 0.6, "grad_norm": 0.0, "learning_rate": 5.81e-06, "loss": 0.017741450667381288, "num_tokens": 1487414.0, "reward": 0.3375, "reward_std": 0.1977891743183136, "rewards/boxed_accuracy_reward/mean": 0.3375, "rewards/boxed_accuracy_reward/std": 0.19778918027877807, "step": 420, "step_time": 25.730258163498366 }, { "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": 358.2, "completions/max_terminated_length": 358.2, "completions/mean_length": 223.7375, "completions/mean_terminated_length": 223.7375, "completions/min_length": 126.5, "completions/min_terminated_length": 126.5, "entropy": 0.27519515678286555, "epoch": 0.043, "frac_reward_zero_std": 0.5, "grad_norm": 0.0, "learning_rate": 5.71e-06, "loss": 0.019644300639629363, "num_tokens": 1518057.0, "reward": 0.1375, "reward_std": 0.19864802658557892, "rewards/boxed_accuracy_reward/mean": 0.1375, "rewards/boxed_accuracy_reward/std": 0.19864802658557892, "step": 430, "step_time": 24.234244189500167 }, { "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": 320.9, "completions/max_terminated_length": 320.9, "completions/mean_length": 207.3625, "completions/mean_terminated_length": 207.3625, "completions/min_length": 140.0, "completions/min_terminated_length": 140.0, "entropy": 0.23536108247935772, "epoch": 0.044, "frac_reward_zero_std": 0.7, "grad_norm": 0.0, "learning_rate": 5.610000000000001e-06, "loss": 0.01632075011730194, "num_tokens": 1549726.0, "reward": 0.2875, "reward_std": 0.13509859144687653, "rewards/boxed_accuracy_reward/mean": 0.2875, "rewards/boxed_accuracy_reward/std": 0.13509859144687653, "step": 440, "step_time": 21.775102456000603 }, { "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": 0.025, "completions/max_length": 491.2, "completions/max_terminated_length": 397.3, "completions/mean_length": 289.6375, "completions/mean_terminated_length": 272.23036041259763, "completions/min_length": 176.8, "completions/min_terminated_length": 176.8, "entropy": 0.20252985879778862, "epoch": 0.045, "frac_reward_zero_std": 0.5, "grad_norm": 0.0, "learning_rate": 5.510000000000001e-06, "loss": 0.07112041711807252, "num_tokens": 1586873.0, "reward": 0.2, "reward_std": 0.23691109120845794, "rewards/boxed_accuracy_reward/mean": 0.2, "rewards/boxed_accuracy_reward/std": 0.23691110908985138, "step": 450, "step_time": 33.158058354499374 }, { "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": 447.2, "completions/max_terminated_length": 447.2, "completions/mean_length": 284.3875, "completions/mean_terminated_length": 284.3875, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "entropy": 0.1773521050810814, "epoch": 0.046, "frac_reward_zero_std": 0.7, "grad_norm": 0.0, "learning_rate": 5.410000000000001e-06, "loss": -0.04271512031555176, "num_tokens": 1622272.0, "reward": 0.25, "reward_std": 0.13340125977993011, "rewards/boxed_accuracy_reward/mean": 0.25, "rewards/boxed_accuracy_reward/std": 0.1334012657403946, "step": 460, "step_time": 30.262695012098995 }, { "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": 0.025, "completions/max_length": 510.3, "completions/max_terminated_length": 417.6, "completions/mean_length": 271.6625, "completions/mean_terminated_length": 254.18214416503906, "completions/min_length": 141.5, "completions/min_terminated_length": 141.5, "entropy": 0.23274324238300323, "epoch": 0.047, "frac_reward_zero_std": 0.5, "grad_norm": 0.0, "learning_rate": 5.310000000000001e-06, "loss": 0.08956294059753418, "num_tokens": 1657333.0, "reward": 0.3, "reward_std": 0.20411193668842315, "rewards/boxed_accuracy_reward/mean": 0.3, "rewards/boxed_accuracy_reward/std": 0.20411194264888763, "step": 470, "step_time": 34.52710425489859 }, { "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": 0.0125, "completions/max_length": 446.0, "completions/max_terminated_length": 396.1, "completions/mean_length": 276.6125, "completions/mean_terminated_length": 268.86607360839844, "completions/min_length": 169.2, "completions/min_terminated_length": 169.2, "entropy": 0.25105012357234957, "epoch": 0.048, "frac_reward_zero_std": 0.8, "grad_norm": 0.0, "learning_rate": 5.210000000000001e-06, "loss": -0.00809616819024086, "num_tokens": 1694526.0, "reward": 0.0375, "reward_std": 0.0816463440656662, "rewards/boxed_accuracy_reward/mean": 0.0375, "rewards/boxed_accuracy_reward/std": 0.0816463440656662, "step": 480, "step_time": 30.273345093600803 }, { "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": 0.0125, "completions/max_length": 442.4, "completions/max_terminated_length": 381.1, "completions/mean_length": 224.5125, "completions/mean_terminated_length": 214.30535736083985, "completions/min_length": 104.2, "completions/min_terminated_length": 104.2, "entropy": 0.24412711150944233, "epoch": 0.049, "frac_reward_zero_std": 0.6, "grad_norm": 0.4479704797267914, "learning_rate": 5.11e-06, "loss": 0.03438442945480347, "num_tokens": 1726799.0, "reward": 0.2125, "reward_std": 0.1759178400039673, "rewards/boxed_accuracy_reward/mean": 0.2125, "rewards/boxed_accuracy_reward/std": 0.17591784596443177, "step": 490, "step_time": 29.914187529801712 }, { "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": 327.1, "completions/max_terminated_length": 327.1, "completions/mean_length": 217.2875, "completions/mean_terminated_length": 217.2875, "completions/min_length": 128.0, "completions/min_terminated_length": 128.0, "entropy": 0.23657882679253817, "epoch": 0.05, "frac_reward_zero_std": 0.6, "grad_norm": 0.6822343468666077, "learning_rate": 5.01e-06, "loss": 0.04140647947788238, "num_tokens": 1758502.0, "reward": 0.275, "reward_std": 0.1632926881313324, "rewards/boxed_accuracy_reward/mean": 0.275, "rewards/boxed_accuracy_reward/std": 0.1632926881313324, "step": 500, "step_time": 22.14599506190134 }, { "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": 296.2, "completions/max_terminated_length": 296.2, "completions/mean_length": 206.925, "completions/mean_terminated_length": 206.925, "completions/min_length": 134.8, "completions/min_terminated_length": 134.8, "entropy": 0.21120237074792386, "epoch": 0.051, "frac_reward_zero_std": 0.4, "grad_norm": 0.0, "learning_rate": 4.9100000000000004e-06, "loss": -0.02729707658290863, "num_tokens": 1787704.0, "reward": 0.4125, "reward_std": 0.2592615157365799, "rewards/boxed_accuracy_reward/mean": 0.4125, "rewards/boxed_accuracy_reward/std": 0.2592615157365799, "step": 510, "step_time": 20.121878111200203 }, { "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": 397.7, "completions/max_terminated_length": 397.7, "completions/mean_length": 263.6125, "completions/mean_terminated_length": 263.6125, "completions/min_length": 164.6, "completions/min_terminated_length": 164.6, "entropy": 0.214491768181324, "epoch": 0.052, "frac_reward_zero_std": 0.6, "grad_norm": 0.0, "learning_rate": 4.8100000000000005e-06, "loss": 0.003942453116178512, "num_tokens": 1822977.0, "reward": 0.0875, "reward_std": 0.16875659823417663, "rewards/boxed_accuracy_reward/mean": 0.0875, "rewards/boxed_accuracy_reward/std": 0.1687566041946411, "step": 520, "step_time": 26.994221128601204 }, { "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": 0.0375, "completions/max_length": 485.0, "completions/max_terminated_length": 358.8, "completions/mean_length": 280.6625, "completions/mean_terminated_length": 254.4261932373047, "completions/min_length": 165.4, "completions/min_terminated_length": 165.4, "entropy": 0.3250377386808395, "epoch": 0.053, "frac_reward_zero_std": 0.6, "grad_norm": 0.6389704346656799, "learning_rate": 4.71e-06, "loss": 0.08574633002281189, "num_tokens": 1859638.0, "reward": 0.15, "reward_std": 0.15782093107700348, "rewards/boxed_accuracy_reward/mean": 0.15, "rewards/boxed_accuracy_reward/std": 0.15782093703746797, "step": 530, "step_time": 32.832222134500626 }, { "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": 361.9, "completions/max_terminated_length": 361.9, "completions/mean_length": 252.375, "completions/mean_terminated_length": 252.375, "completions/min_length": 164.6, "completions/min_terminated_length": 164.6, "entropy": 0.2565031200647354, "epoch": 0.054, "frac_reward_zero_std": 0.6, "grad_norm": 0.0, "learning_rate": 4.610000000000001e-06, "loss": 0.021490342915058136, "num_tokens": 1893300.0, "reward": 0.1625, "reward_std": 0.20871699452400208, "rewards/boxed_accuracy_reward/mean": 0.1625, "rewards/boxed_accuracy_reward/std": 0.2087170124053955, "step": 540, "step_time": 24.475655020699925 }, { "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": 0.025, "completions/max_length": 373.8, "completions/max_terminated_length": 299.6, "completions/mean_length": 206.3, "completions/mean_terminated_length": 185.17916717529297, "completions/min_length": 111.3, "completions/min_terminated_length": 111.3, "entropy": 0.22958396673202514, "epoch": 0.055, "frac_reward_zero_std": 0.3, "grad_norm": 0.6256431937217712, "learning_rate": 4.510000000000001e-06, "loss": 0.03615856468677521, "num_tokens": 1923932.0, "reward": 0.375, "reward_std": 0.29838343262672423, "rewards/boxed_accuracy_reward/mean": 0.375, "rewards/boxed_accuracy_reward/std": 0.2983834445476532, "step": 550, "step_time": 25.349692796901216 }, { "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": 0.05, "completions/max_length": 475.5, "completions/max_terminated_length": 364.6, "completions/mean_length": 278.4375, "completions/mean_terminated_length": 239.52916870117187, "completions/min_length": 158.6, "completions/min_terminated_length": 158.6, "entropy": 0.2459364876151085, "epoch": 0.056, "frac_reward_zero_std": 0.5, "grad_norm": 0.8983960151672363, "learning_rate": 4.41e-06, "loss": -0.031027239561080933, "num_tokens": 1963111.0, "reward": 0.075, "reward_std": 0.18771235942840575, "rewards/boxed_accuracy_reward/mean": 0.075, "rewards/boxed_accuracy_reward/std": 0.18771235942840575, "step": 560, "step_time": 32.03998489129954 }, { "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": 0.0125, "completions/max_length": 419.6, "completions/max_terminated_length": 378.6, "completions/mean_length": 257.3625, "completions/mean_terminated_length": 249.33035888671876, "completions/min_length": 171.2, "completions/min_terminated_length": 171.2, "entropy": 0.2526143785566092, "epoch": 0.057, "frac_reward_zero_std": 0.7, "grad_norm": 0.0, "learning_rate": 4.31e-06, "loss": 0.01212334781885147, "num_tokens": 1998908.0, "reward": 0.2, "reward_std": 0.14056250154972078, "rewards/boxed_accuracy_reward/mean": 0.2, "rewards/boxed_accuracy_reward/std": 0.14056250751018523, "step": 570, "step_time": 28.352883074600687 }, { "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": 394.2, "completions/max_terminated_length": 394.2, "completions/mean_length": 257.6875, "completions/mean_terminated_length": 257.6875, "completions/min_length": 160.4, "completions/min_terminated_length": 160.4, "entropy": 0.2496233806014061, "epoch": 0.058, "frac_reward_zero_std": 0.5, "grad_norm": 0.6807348132133484, "learning_rate": 4.21e-06, "loss": 0.0322126567363739, "num_tokens": 2034387.0, "reward": 0.1375, "reward_std": 0.1767766922712326, "rewards/boxed_accuracy_reward/mean": 0.1375, "rewards/boxed_accuracy_reward/std": 0.1767766922712326, "step": 580, "step_time": 26.78779618450062 }, { "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": 0.0125, "completions/max_length": 371.7, "completions/max_terminated_length": 302.6, "completions/mean_length": 208.2625, "completions/mean_terminated_length": 198.36964416503906, "completions/min_length": 121.9, "completions/min_terminated_length": 121.9, "entropy": 0.22029370740056037, "epoch": 0.059, "frac_reward_zero_std": 0.6, "grad_norm": 0.0, "learning_rate": 4.1100000000000005e-06, "loss": 0.0259624183177948, "num_tokens": 2067680.0, "reward": 0.275, "reward_std": 0.18685350716114044, "rewards/boxed_accuracy_reward/mean": 0.275, "rewards/boxed_accuracy_reward/std": 0.18685351312160492, "step": 590, "step_time": 25.24782887190013 }, { "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": 0.025, "completions/max_length": 447.7, "completions/max_terminated_length": 380.5, "completions/mean_length": 247.1375, "completions/mean_terminated_length": 229.69286193847657, "completions/min_length": 137.8, "completions/min_terminated_length": 137.8, "entropy": 0.24857322350144387, "epoch": 0.06, "frac_reward_zero_std": 0.7, "grad_norm": 0.0, "learning_rate": 4.0100000000000006e-06, "loss": 0.04133725166320801, "num_tokens": 2101531.0, "reward": 0.1, "reward_std": 0.13340125977993011, "rewards/boxed_accuracy_reward/mean": 0.1, "rewards/boxed_accuracy_reward/std": 0.1334012657403946, "step": 600, "step_time": 30.182934339900385 }, { "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": 0.0125, "completions/max_length": 423.4, "completions/max_terminated_length": 362.5, "completions/mean_length": 264.475, "completions/mean_terminated_length": 255.80714416503906, "completions/min_length": 171.7, "completions/min_terminated_length": 171.7, "entropy": 0.23762509562075138, "epoch": 0.061, "frac_reward_zero_std": 0.5, "grad_norm": 0.8978026509284973, "learning_rate": 3.910000000000001e-06, "loss": 0.04321916103363037, "num_tokens": 2137113.0, "reward": 0.4, "reward_std": 0.24030575454235076, "rewards/boxed_accuracy_reward/mean": 0.4, "rewards/boxed_accuracy_reward/std": 0.24030576050281524, "step": 610, "step_time": 28.61728952409976 }, { "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": 368.3, "completions/max_terminated_length": 368.3, "completions/mean_length": 232.825, "completions/mean_terminated_length": 232.825, "completions/min_length": 147.5, "completions/min_terminated_length": 147.5, "entropy": 0.2583306774497032, "epoch": 0.062, "frac_reward_zero_std": 0.7, "grad_norm": 0.4808867573738098, "learning_rate": 3.8100000000000004e-06, "loss": 0.0032923467457294463, "num_tokens": 2169507.0, "reward": 0.0625, "reward_std": 0.12793734967708587, "rewards/boxed_accuracy_reward/mean": 0.0625, "rewards/boxed_accuracy_reward/std": 0.12793734967708587, "step": 620, "step_time": 24.657616854002118 }, { "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": 393.6, "completions/max_terminated_length": 393.6, "completions/mean_length": 239.7875, "completions/mean_terminated_length": 239.7875, "completions/min_length": 150.1, "completions/min_terminated_length": 150.1, "entropy": 0.24895308986306192, "epoch": 0.063, "frac_reward_zero_std": 0.6, "grad_norm": 0.0, "learning_rate": 3.7100000000000005e-06, "loss": -0.008826246112585067, "num_tokens": 2202426.0, "reward": 0.1625, "reward_std": 0.19062793254852295, "rewards/boxed_accuracy_reward/mean": 0.1625, "rewards/boxed_accuracy_reward/std": 0.19062793850898743, "step": 630, "step_time": 26.481925834499272 }, { "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": 0.0125, "completions/max_length": 490.3, "completions/max_terminated_length": 412.0, "completions/mean_length": 271.6625, "completions/mean_terminated_length": 261.0321441650391, "completions/min_length": 156.2, "completions/min_terminated_length": 156.2, "entropy": 0.2555754192173481, "epoch": 0.064, "frac_reward_zero_std": 0.7, "grad_norm": 0.0, "learning_rate": 3.61e-06, "loss": -0.015103229880332946, "num_tokens": 2237311.0, "reward": 0.225, "reward_std": 0.13340125977993011, "rewards/boxed_accuracy_reward/mean": 0.225, "rewards/boxed_accuracy_reward/std": 0.1334012657403946, "step": 640, "step_time": 32.94274350120031 }, { "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": 445.9, "completions/max_terminated_length": 445.9, "completions/mean_length": 247.9125, "completions/mean_terminated_length": 247.9125, "completions/min_length": 134.0, "completions/min_terminated_length": 134.0, "entropy": 0.28245309442281724, "epoch": 0.065, "frac_reward_zero_std": 0.6, "grad_norm": 0.0, "learning_rate": 3.5100000000000003e-06, "loss": 0.05236095190048218, "num_tokens": 2271440.0, "reward": 0.15, "reward_std": 0.18685350716114044, "rewards/boxed_accuracy_reward/mean": 0.15, "rewards/boxed_accuracy_reward/std": 0.18685351312160492, "step": 650, "step_time": 30.18701468149884 }, { "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": 0.0125, "completions/max_length": 364.8, "completions/max_terminated_length": 317.3, "completions/mean_length": 223.35, "completions/mean_terminated_length": 214.17857360839844, "completions/min_length": 135.9, "completions/min_terminated_length": 135.9, "entropy": 0.265711085498333, "epoch": 0.066, "frac_reward_zero_std": 0.3, "grad_norm": 0.8590646982192993, "learning_rate": 3.4100000000000004e-06, "loss": 0.03245477676391602, "num_tokens": 2304332.0, "reward": 0.3375, "reward_std": 0.3311904340982437, "rewards/boxed_accuracy_reward/mean": 0.3375, "rewards/boxed_accuracy_reward/std": 0.3311904460191727, "step": 660, "step_time": 24.758795369901055 }, { "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": 369.7, "completions/max_terminated_length": 369.7, "completions/mean_length": 255.3375, "completions/mean_terminated_length": 255.3375, "completions/min_length": 146.1, "completions/min_terminated_length": 146.1, "entropy": 0.2967137709259987, "epoch": 0.067, "frac_reward_zero_std": 0.8, "grad_norm": 0.8069291710853577, "learning_rate": 3.3100000000000005e-06, "loss": 0.022205421328544618, "num_tokens": 2338295.0, "reward": 0.2, "reward_std": 0.10350983142852783, "rewards/boxed_accuracy_reward/mean": 0.2, "rewards/boxed_accuracy_reward/std": 0.10350984334945679, "step": 670, "step_time": 24.84447250970188 }, { "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": 378.8, "completions/max_terminated_length": 378.8, "completions/mean_length": 223.475, "completions/mean_terminated_length": 223.475, "completions/min_length": 140.8, "completions/min_terminated_length": 140.8, "entropy": 0.3041005529463291, "epoch": 0.068, "frac_reward_zero_std": 0.5, "grad_norm": 0.0, "learning_rate": 3.21e-06, "loss": -0.020687612891197204, "num_tokens": 2369605.0, "reward": 0.2, "reward_std": 0.24954409003257752, "rewards/boxed_accuracy_reward/mean": 0.2, "rewards/boxed_accuracy_reward/std": 0.24954410195350646, "step": 680, "step_time": 25.69439109330124 }, { "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": 346.6, "completions/max_terminated_length": 346.6, "completions/mean_length": 237.9625, "completions/mean_terminated_length": 237.9625, "completions/min_length": 147.4, "completions/min_terminated_length": 147.4, "entropy": 0.24550328701734542, "epoch": 0.069, "frac_reward_zero_std": 0.7, "grad_norm": 0.9554150700569153, "learning_rate": 3.1100000000000003e-06, "loss": 0.013637937605381012, "num_tokens": 2401898.0, "reward": 0.1875, "reward_std": 0.13509859144687653, "rewards/boxed_accuracy_reward/mean": 0.1875, "rewards/boxed_accuracy_reward/std": 0.13509859144687653, "step": 690, "step_time": 23.408739431599678 }, { "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": 390.2, "completions/max_terminated_length": 390.2, "completions/mean_length": 242.8125, "completions/mean_terminated_length": 242.8125, "completions/min_length": 145.6, "completions/min_terminated_length": 145.6, "entropy": 0.2589107438921928, "epoch": 0.07, "frac_reward_zero_std": 0.6, "grad_norm": 1.1996217966079712, "learning_rate": 3.01e-06, "loss": 0.05301226377487182, "num_tokens": 2436907.0, "reward": 0.1125, "reward_std": 0.18138959705829621, "rewards/boxed_accuracy_reward/mean": 0.1125, "rewards/boxed_accuracy_reward/std": 0.18138959705829621, "step": 700, "step_time": 26.322794716500358 }, { "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": 0.0125, "completions/max_length": 369.9, "completions/max_terminated_length": 291.4, "completions/mean_length": 211.5125, "completions/mean_terminated_length": 200.55178680419922, "completions/min_length": 133.0, "completions/min_terminated_length": 133.0, "entropy": 0.19192061349749565, "epoch": 0.071, "frac_reward_zero_std": 0.9, "grad_norm": 0.0, "learning_rate": 2.91e-06, "loss": -0.003464340791106224, "num_tokens": 2467292.0, "reward": 0.2875, "reward_std": 0.03535533845424652, "rewards/boxed_accuracy_reward/mean": 0.2875, "rewards/boxed_accuracy_reward/std": 0.03535533845424652, "step": 710, "step_time": 25.09874547720101 }, { "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": 0.05, "completions/max_length": 574.9, "completions/max_terminated_length": 388.2, "completions/mean_length": 293.5125, "completions/mean_terminated_length": 260.60000762939455, "completions/min_length": 169.5, "completions/min_terminated_length": 169.5, "entropy": 0.22459689602255822, "epoch": 0.072, "frac_reward_zero_std": 0.6, "grad_norm": 0.2952307164669037, "learning_rate": 2.8100000000000006e-06, "loss": 0.04532091617584229, "num_tokens": 2502805.0, "reward": 0.25, "reward_std": 0.15782093107700348, "rewards/boxed_accuracy_reward/mean": 0.25, "rewards/boxed_accuracy_reward/std": 0.15782093703746797, "step": 720, "step_time": 38.782466992000266 }, { "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": 315.9, "completions/max_terminated_length": 315.9, "completions/mean_length": 216.875, "completions/mean_terminated_length": 216.875, "completions/min_length": 130.2, "completions/min_terminated_length": 130.2, "entropy": 0.26642010658979415, "epoch": 0.073, "frac_reward_zero_std": 0.6, "grad_norm": 0.7213253974914551, "learning_rate": 2.7100000000000003e-06, "loss": -0.004600425064563751, "num_tokens": 2535131.0, "reward": 0.1375, "reward_std": 0.17422835528850555, "rewards/boxed_accuracy_reward/mean": 0.1375, "rewards/boxed_accuracy_reward/std": 0.17422835528850555, "step": 730, "step_time": 21.296572963699145 }, { "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": 427.0, "completions/max_terminated_length": 427.0, "completions/mean_length": 246.8375, "completions/mean_terminated_length": 246.8375, "completions/min_length": 159.4, "completions/min_terminated_length": 159.4, "entropy": 0.2320882935076952, "epoch": 0.074, "frac_reward_zero_std": 0.6, "grad_norm": 0.0, "learning_rate": 2.6100000000000004e-06, "loss": -0.0016866859048604965, "num_tokens": 2567462.0, "reward": 0.225, "reward_std": 0.1796922653913498, "rewards/boxed_accuracy_reward/mean": 0.225, "rewards/boxed_accuracy_reward/std": 0.17969227135181426, "step": 740, "step_time": 28.78491548419988 }, { "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": 0.0125, "completions/max_length": 407.3, "completions/max_terminated_length": 341.9, "completions/mean_length": 229.175, "completions/mean_terminated_length": 218.69285736083984, "completions/min_length": 129.1, "completions/min_terminated_length": 129.1, "entropy": 0.2695804834365845, "epoch": 0.075, "frac_reward_zero_std": 0.6, "grad_norm": 1.118378758430481, "learning_rate": 2.51e-06, "loss": 0.02082826942205429, "num_tokens": 2601372.0, "reward": 0.175, "reward_std": 0.17045392990112304, "rewards/boxed_accuracy_reward/mean": 0.175, "rewards/boxed_accuracy_reward/std": 0.17045392990112304, "step": 750, "step_time": 27.563064585998653 }, { "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": 250.8, "completions/max_terminated_length": 250.8, "completions/mean_length": 172.0375, "completions/mean_terminated_length": 172.0375, "completions/min_length": 121.6, "completions/min_terminated_length": 121.6, "entropy": 0.2495213456451893, "epoch": 0.076, "frac_reward_zero_std": 0.7, "grad_norm": 1.1198997497558594, "learning_rate": 2.4100000000000002e-06, "loss": -0.0004969652742147446, "num_tokens": 2631303.0, "reward": 0.3125, "reward_std": 0.14433692693710326, "rewards/boxed_accuracy_reward/mean": 0.3125, "rewards/boxed_accuracy_reward/std": 0.14433693289756774, "step": 760, "step_time": 17.046090553899557 }, { "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": 341.0, "completions/max_terminated_length": 341.0, "completions/mean_length": 209.725, "completions/mean_terminated_length": 209.725, "completions/min_length": 120.9, "completions/min_terminated_length": 120.9, "entropy": 0.2746701970696449, "epoch": 0.077, "frac_reward_zero_std": 0.5, "grad_norm": 0.0, "learning_rate": 2.3100000000000003e-06, "loss": 0.04669084250926971, "num_tokens": 2661489.0, "reward": 0.25, "reward_std": 0.23691109120845794, "rewards/boxed_accuracy_reward/mean": 0.25, "rewards/boxed_accuracy_reward/std": 0.23691110908985138, "step": 770, "step_time": 23.079953304801894 }, { "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": 0.0625, "completions/max_length": 599.1, "completions/max_terminated_length": 445.5, "completions/mean_length": 323.7625, "completions/mean_terminated_length": 287.5857177734375, "completions/min_length": 165.4, "completions/min_terminated_length": 165.4, "entropy": 0.20161464884877206, "epoch": 0.078, "frac_reward_zero_std": 0.7, "grad_norm": 0.6654395461082458, "learning_rate": 2.21e-06, "loss": -0.01412966102361679, "num_tokens": 2703534.0, "reward": 0.2, "reward_std": 0.14056250154972078, "rewards/boxed_accuracy_reward/mean": 0.2, "rewards/boxed_accuracy_reward/std": 0.14056250751018523, "step": 780, "step_time": 40.28982557420022 }, { "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": 0.025, "completions/max_length": 488.7, "completions/max_terminated_length": 422.1, "completions/mean_length": 281.3625, "completions/mean_terminated_length": 261.6958343505859, "completions/min_length": 144.9, "completions/min_terminated_length": 144.9, "entropy": 0.23346679359674455, "epoch": 0.079, "frac_reward_zero_std": 0.5, "grad_norm": 0.0, "learning_rate": 2.11e-06, "loss": 0.03522632122039795, "num_tokens": 2739035.0, "reward": 0.25, "reward_std": 0.2529387533664703, "rewards/boxed_accuracy_reward/mean": 0.25, "rewards/boxed_accuracy_reward/std": 0.2529387533664703, "step": 790, "step_time": 33.08498501170034 }, { "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": 344.0, "completions/max_terminated_length": 344.0, "completions/mean_length": 210.85, "completions/mean_terminated_length": 210.85, "completions/min_length": 133.1, "completions/min_terminated_length": 133.1, "entropy": 0.25274474173784256, "epoch": 0.08, "frac_reward_zero_std": 0.7, "grad_norm": 0.33814314007759094, "learning_rate": 2.0100000000000002e-06, "loss": 0.009773983806371688, "num_tokens": 2768951.0, "reward": 0.2, "reward_std": 0.13340125977993011, "rewards/boxed_accuracy_reward/mean": 0.2, "rewards/boxed_accuracy_reward/std": 0.1334012657403946, "step": 800, "step_time": 23.216696536900418 } ], "logging_steps": 10, "max_steps": 1000, "num_input_tokens_seen": 2768951, "num_train_epochs": 1, "save_steps": 200, "stateful_callbacks": { "TrainerControl": { "args": { "should_epoch_stop": false, "should_evaluate": false, "should_log": false, "should_save": true, "should_training_stop": false }, "attributes": {} } }, "total_flos": 0.0, "train_batch_size": 8, "trial_name": null, "trial_params": null }