{ "best_global_step": null, "best_metric": null, "best_model_checkpoint": null, "epoch": 0.25, "eval_steps": 500, "global_step": 441, "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3204.0, "completions/mean_length": 1194.96875, "completions/mean_terminated_length": 894.862060546875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.504598390311003, "epoch": 0.0005668934240362812, "frac_reward_zero_std": 0.75, "grad_norm": 0.189453125, "learning_rate": 1e-06, "loss": 0.0204, "num_tokens": 86206.0, "reward": 0.1015625, "reward_std": 0.40999117493629456, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 1, "step_time": 288.6456542061642 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3661.0, "completions/mean_length": 1249.453125, "completions/mean_terminated_length": 954.9827270507812, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.565896205604076, "epoch": 0.0011337868480725624, "frac_reward_zero_std": 0.5, "grad_norm": 0.265625, "learning_rate": 9.994331065759637e-07, "loss": 0.0236, "num_tokens": 179163.0, "reward": 0.3046875, "reward_std": 0.6878606677055359, "rewards/accuracy_reward_func/mean": 0.25, "rewards/accuracy_reward_func/std": 0.6666666865348816, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 2, "step_time": 301.0371378036216 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3855.0, "completions/mean_length": 1619.375, "completions/mean_terminated_length": 1047.84619140625, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 4.525065306574106, "epoch": 0.0017006802721088435, "frac_reward_zero_std": 0.75, "grad_norm": 0.1708984375, "learning_rate": 9.988662131519275e-07, "loss": 0.0446, "num_tokens": 293507.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 3, "step_time": 295.80338739976287 }, { "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": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 2797.0, "completions/mean_length": 1325.203125, "completions/mean_terminated_length": 618.9215698242188, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.062012603506446, "epoch": 0.0022675736961451248, "frac_reward_zero_std": 0.625, "grad_norm": 0.2470703125, "learning_rate": 9.982993197278912e-07, "loss": 0.0291, "num_tokens": 389872.0, "reward": 0.109375, "reward_std": 0.37234508991241455, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 4, "step_time": 290.4045874308795 }, { "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": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4050.0, "completions/mean_length": 1660.046875, "completions/mean_terminated_length": 977.97998046875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.120975449681282, "epoch": 0.002834467120181406, "frac_reward_zero_std": 0.875, "grad_norm": 0.1484375, "learning_rate": 9.977324263038547e-07, "loss": -0.0061, "num_tokens": 507171.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 5, "step_time": 290.6480199228972 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3879.0, "completions/mean_length": 1265.65625, "completions/mean_terminated_length": 918.0701904296875, "completions/min_length": 29.0, "completions/min_terminated_length": 29.0, "entropy": 3.452789928764105, "epoch": 0.003401360544217687, "frac_reward_zero_std": 0.875, "grad_norm": 0.216796875, "learning_rate": 9.971655328798185e-07, "loss": -0.0297, "num_tokens": 600277.0, "reward": 0.03125, "reward_std": 0.25, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 6, "step_time": 293.04367576539516 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3573.0, "completions/mean_length": 1149.5625, "completions/mean_terminated_length": 787.7192993164062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.292115245014429, "epoch": 0.003968253968253968, "frac_reward_zero_std": 0.75, "grad_norm": 0.1826171875, "learning_rate": 9.965986394557822e-07, "loss": 0.0167, "num_tokens": 687833.0, "reward": 0.1484375, "reward_std": 0.5092765688896179, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 7, "step_time": 306.0402205856517 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4072.0, "completions/mean_length": 1350.578125, "completions/mean_terminated_length": 958.3750610351562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.9101115502417088, "epoch": 0.0045351473922902496, "frac_reward_zero_std": 0.875, "grad_norm": 0.171875, "learning_rate": 9.96031746031746e-07, "loss": 0.0391, "num_tokens": 787358.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 8, "step_time": 294.3254048600793 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3816.0, "completions/mean_length": 1264.828125, "completions/mean_terminated_length": 917.140380859375, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "entropy": 2.6403061226010323, "epoch": 0.00510204081632653, "frac_reward_zero_std": 0.5, "grad_norm": 0.30859375, "learning_rate": 9.954648526077097e-07, "loss": 0.0312, "num_tokens": 880947.0, "reward": 0.1875, "reward_std": 0.5739033818244934, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 9, "step_time": 295.80457826238126 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3332.0, "completions/mean_length": 1584.625, "completions/mean_terminated_length": 1005.0769653320312, "completions/min_length": 198.0, "completions/min_terminated_length": 198.0, "entropy": 4.103184837847948, "epoch": 0.005668934240362812, "frac_reward_zero_std": 0.875, "grad_norm": 0.11669921875, "learning_rate": 9.948979591836735e-07, "loss": 0.001, "num_tokens": 992979.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 10, "step_time": 292.4153257710859 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3863.0, "completions/mean_length": 1550.484375, "completions/mean_terminated_length": 1079.0926513671875, "completions/min_length": 110.0, "completions/min_terminated_length": 110.0, "entropy": 3.6203736774623394, "epoch": 0.006235827664399093, "frac_reward_zero_std": 0.75, "grad_norm": 0.25, "learning_rate": 9.94331065759637e-07, "loss": -0.0192, "num_tokens": 1103218.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 11, "step_time": 293.3653084374964 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3629.0, "completions/mean_length": 977.4375, "completions/mean_terminated_length": 824.0655517578125, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "entropy": 3.601443100720644, "epoch": 0.006802721088435374, "frac_reward_zero_std": 0.5, "grad_norm": 0.4296875, "learning_rate": 9.937641723356008e-07, "loss": -0.0671, "num_tokens": 1177038.0, "reward": 0.1640625, "reward_std": 0.496453195810318, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 12, "step_time": 291.27144080307335 }, { "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": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3930.0, "completions/mean_length": 1514.34375, "completions/mean_terminated_length": 856.2745361328125, "completions/min_length": 19.0, "completions/min_terminated_length": 19.0, "entropy": 3.1219849698245525, "epoch": 0.007369614512471655, "frac_reward_zero_std": 0.875, "grad_norm": 0.10693359375, "learning_rate": 9.931972789115645e-07, "loss": 0.0084, "num_tokens": 1286332.0, "reward": 0.078125, "reward_std": 0.4005824029445648, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 13, "step_time": 291.7519850637764 }, { "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": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4057.0, "completions/mean_length": 1760.421875, "completions/mean_terminated_length": 1106.4599609375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.9482052512466908, "epoch": 0.007936507936507936, "frac_reward_zero_std": 0.75, "grad_norm": 0.1328125, "learning_rate": 9.926303854875283e-07, "loss": 0.0131, "num_tokens": 1411599.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 14, "step_time": 294.686209958978 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3995.0, "completions/mean_length": 1343.328125, "completions/mean_terminated_length": 1005.2807006835938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.780040428042412, "epoch": 0.008503401360544218, "frac_reward_zero_std": 0.75, "grad_norm": 0.1826171875, "learning_rate": 9.92063492063492e-07, "loss": 0.0169, "num_tokens": 1509612.0, "reward": 0.1875, "reward_std": 0.6267831921577454, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 15, "step_time": 291.73783070314676 }, { "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": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3489.0, "completions/mean_length": 1590.0625, "completions/mean_terminated_length": 951.2941284179688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.6598825938999653, "epoch": 0.009070294784580499, "frac_reward_zero_std": 0.75, "grad_norm": 0.1767578125, "learning_rate": 9.914965986394558e-07, "loss": 0.0232, "num_tokens": 1622576.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 16, "step_time": 293.3912907987833 }, { "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": 0.296875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3930.0, "completions/mean_length": 2036.125, "completions/mean_terminated_length": 1166.4000244140625, "completions/min_length": 21.0, "completions/min_terminated_length": 21.0, "entropy": 5.264551121741533, "epoch": 0.00963718820861678, "frac_reward_zero_std": 0.75, "grad_norm": 0.1328125, "learning_rate": 9.909297052154193e-07, "loss": 0.0406, "num_tokens": 1765160.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 17, "step_time": 292.5890276953578 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3973.0, "completions/mean_length": 1329.9375, "completions/mean_terminated_length": 1043.7930908203125, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "entropy": 4.571435987949371, "epoch": 0.01020408163265306, "frac_reward_zero_std": 0.875, "grad_norm": 0.162109375, "learning_rate": 9.90362811791383e-07, "loss": 0.0015, "num_tokens": 1861252.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 18, "step_time": 292.55902184545994 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4078.0, "completions/mean_length": 927.75, "completions/mean_terminated_length": 659.2542114257812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.181705266237259, "epoch": 0.010770975056689343, "frac_reward_zero_std": 0.75, "grad_norm": 0.294921875, "learning_rate": 9.897959183673468e-07, "loss": 0.0149, "num_tokens": 1931516.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 19, "step_time": 289.3771469090134 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3936.0, "completions/mean_length": 1601.03125, "completions/mean_terminated_length": 1083.20751953125, "completions/min_length": 161.0, "completions/min_terminated_length": 161.0, "entropy": 4.274967905133963, "epoch": 0.011337868480725623, "frac_reward_zero_std": 0.75, "grad_norm": 0.1142578125, "learning_rate": 9.892290249433106e-07, "loss": 0.0122, "num_tokens": 2045406.0, "reward": 0.0625, "reward_std": 0.35073620080947876, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 20, "step_time": 298.2229356933385 }, { "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": 0.28125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4001.0, "completions/mean_length": 1762.90625, "completions/mean_terminated_length": 849.95654296875, "completions/min_length": 89.0, "completions/min_terminated_length": 89.0, "entropy": 2.933218887075782, "epoch": 0.011904761904761904, "frac_reward_zero_std": 0.875, "grad_norm": 0.1220703125, "learning_rate": 9.886621315192743e-07, "loss": 0.0323, "num_tokens": 2172472.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 21, "step_time": 300.25128427147865 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3950.0, "completions/mean_length": 1170.28125, "completions/mean_terminated_length": 867.6206665039062, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 3.5906212478876114, "epoch": 0.012471655328798186, "frac_reward_zero_std": 0.75, "grad_norm": 0.1728515625, "learning_rate": 9.88095238095238e-07, "loss": 0.0181, "num_tokens": 2267298.0, "reward": 0.125, "reward_std": 0.417855441570282, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 22, "step_time": 339.2186377122998 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3624.0, "completions/mean_length": 1314.890625, "completions/mean_terminated_length": 859.7999877929688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.1952144782990217, "epoch": 0.013038548752834467, "frac_reward_zero_std": 0.625, "grad_norm": 0.1865234375, "learning_rate": 9.875283446712018e-07, "loss": -0.0041, "num_tokens": 2363667.0, "reward": 0.1640625, "reward_std": 0.5274578928947449, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 23, "step_time": 292.4616867406294 }, { "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": 0.28125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4027.0, "completions/mean_length": 1822.078125, "completions/mean_terminated_length": 932.2826538085938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.461105719208717, "epoch": 0.013605442176870748, "frac_reward_zero_std": 0.875, "grad_norm": 0.115234375, "learning_rate": 9.869614512471656e-07, "loss": 0.0031, "num_tokens": 2492088.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 24, "step_time": 292.1666828431189 }, { "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": 4096.0, "completions/max_terminated_length": 2464.0, "completions/mean_length": 972.109375, "completions/mean_terminated_length": 763.8500366210938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.696283109486103, "epoch": 0.01417233560090703, "frac_reward_zero_std": 0.75, "grad_norm": 0.271484375, "learning_rate": 9.86394557823129e-07, "loss": 0.033, "num_tokens": 2565815.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 25, "step_time": 294.38425650354475 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3978.0, "completions/mean_length": 1340.859375, "completions/mean_terminated_length": 830.6481323242188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.3921100050210953, "epoch": 0.01473922902494331, "frac_reward_zero_std": 0.875, "grad_norm": 0.125, "learning_rate": 9.858276643990928e-07, "loss": 0.0106, "num_tokens": 2662222.0, "reward": 0.0546875, "reward_std": 0.32265254855155945, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 26, "step_time": 289.9887812053785 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3897.0, "completions/mean_length": 1486.921875, "completions/mean_terminated_length": 1059.9818115234375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.797033675014973, "epoch": 0.015306122448979591, "frac_reward_zero_std": 0.75, "grad_norm": 0.17578125, "learning_rate": 9.852607709750566e-07, "loss": 0.0283, "num_tokens": 2769241.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 27, "step_time": 290.794472405687 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3697.0, "completions/mean_length": 970.265625, "completions/mean_terminated_length": 705.3728637695312, "completions/min_length": 96.0, "completions/min_terminated_length": 96.0, "entropy": 3.3921418748795986, "epoch": 0.015873015873015872, "frac_reward_zero_std": 0.625, "grad_norm": 0.25390625, "learning_rate": 9.846938775510203e-07, "loss": 0.0207, "num_tokens": 2842466.0, "reward": 0.1796875, "reward_std": 0.5729843378067017, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 28, "step_time": 291.1961542898789 }, { "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": 0.265625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3858.0, "completions/mean_length": 1910.234375, "completions/mean_terminated_length": 1119.6383056640625, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "entropy": 3.4209898710250854, "epoch": 0.016439909297052153, "frac_reward_zero_std": 0.875, "grad_norm": 0.11669921875, "learning_rate": 9.84126984126984e-07, "loss": 0.0164, "num_tokens": 2978977.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 29, "step_time": 296.4940730072558 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3934.0, "completions/mean_length": 1537.609375, "completions/mean_terminated_length": 1006.6226806640625, "completions/min_length": 17.0, "completions/min_terminated_length": 17.0, "entropy": 3.9501586817204952, "epoch": 0.017006802721088437, "frac_reward_zero_std": 0.875, "grad_norm": 0.126953125, "learning_rate": 9.835600907029478e-07, "loss": 0.0013, "num_tokens": 3090040.0, "reward": 0.046875, "reward_std": 0.3177144229412079, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 30, "step_time": 302.78896435908973 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4016.0, "completions/mean_length": 1483.109375, "completions/mean_terminated_length": 940.8113403320312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.6506435722112656, "epoch": 0.017573696145124718, "frac_reward_zero_std": 0.625, "grad_norm": 0.2294921875, "learning_rate": 9.829931972789116e-07, "loss": 0.0592, "num_tokens": 3195895.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 31, "step_time": 292.1012362819165 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3849.0, "completions/mean_length": 1257.921875, "completions/mean_terminated_length": 964.3275756835938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.385785397142172, "epoch": 0.018140589569160998, "frac_reward_zero_std": 0.75, "grad_norm": 0.2119140625, "learning_rate": 9.824263038548753e-07, "loss": 0.0033, "num_tokens": 3290258.0, "reward": 0.1328125, "reward_std": 0.4386323094367981, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 32, "step_time": 302.2366702305153 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4001.0, "completions/mean_length": 1520.28125, "completions/mean_terminated_length": 985.6981201171875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.249439466744661, "epoch": 0.01870748299319728, "frac_reward_zero_std": 0.625, "grad_norm": 0.208984375, "learning_rate": 9.818594104308389e-07, "loss": 0.0348, "num_tokens": 3397484.0, "reward": 0.2421875, "reward_std": 0.629891574382782, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 33, "step_time": 291.45262563135475 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3852.0, "completions/mean_length": 1365.9375, "completions/mean_terminated_length": 1083.5172119140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.4336040429770947, "epoch": 0.01927437641723356, "frac_reward_zero_std": 0.75, "grad_norm": 0.185546875, "learning_rate": 9.812925170068026e-07, "loss": 0.0323, "num_tokens": 3496336.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 34, "step_time": 290.97266877349466 }, { "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": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3990.0, "completions/mean_length": 1852.953125, "completions/mean_terminated_length": 1224.9000244140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.439823977649212, "epoch": 0.01984126984126984, "frac_reward_zero_std": 0.625, "grad_norm": 0.19140625, "learning_rate": 9.807256235827664e-07, "loss": 0.0135, "num_tokens": 3627781.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 35, "step_time": 298.9179294956848 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3879.0, "completions/mean_length": 1549.28125, "completions/mean_terminated_length": 1132.54541015625, "completions/min_length": 36.0, "completions/min_terminated_length": 36.0, "entropy": 4.371560100466013, "epoch": 0.02040816326530612, "frac_reward_zero_std": 0.875, "grad_norm": 0.1513671875, "learning_rate": 9.801587301587301e-07, "loss": 0.0239, "num_tokens": 3739063.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 36, "step_time": 295.9237609980628 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3804.0, "completions/mean_length": 1154.75, "completions/mean_terminated_length": 793.5438842773438, "completions/min_length": 36.0, "completions/min_terminated_length": 36.0, "entropy": 3.133853852748871, "epoch": 0.020975056689342405, "frac_reward_zero_std": 0.625, "grad_norm": 0.240234375, "learning_rate": 9.795918367346939e-07, "loss": 0.0391, "num_tokens": 3825663.0, "reward": 0.1796875, "reward_std": 0.5445777773857117, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 37, "step_time": 300.0141163431108 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4011.0, "completions/mean_length": 1362.140625, "completions/mean_terminated_length": 971.58935546875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.014933530241251, "epoch": 0.021541950113378686, "frac_reward_zero_std": 0.75, "grad_norm": 0.220703125, "learning_rate": 9.790249433106576e-07, "loss": 0.0314, "num_tokens": 3923648.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 38, "step_time": 289.4139108415693 }, { "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": 0.265625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3834.0, "completions/mean_length": 1936.890625, "completions/mean_terminated_length": 1155.9361572265625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "entropy": 5.069778256118298, "epoch": 0.022108843537414966, "frac_reward_zero_std": 0.875, "grad_norm": 0.1162109375, "learning_rate": 9.784580498866214e-07, "loss": 0.0153, "num_tokens": 4059873.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 39, "step_time": 292.76254057046026 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3938.0, "completions/mean_length": 1399.109375, "completions/mean_terminated_length": 1120.1207275390625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.864141821861267, "epoch": 0.022675736961451247, "frac_reward_zero_std": 0.625, "grad_norm": 0.287109375, "learning_rate": 9.778911564625851e-07, "loss": -0.036, "num_tokens": 4165328.0, "reward": 0.125, "reward_std": 0.4364357888698578, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 40, "step_time": 311.8287929845974 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3994.0, "completions/mean_length": 1269.96875, "completions/mean_terminated_length": 807.5272216796875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.2809007838368416, "epoch": 0.023242630385487528, "frac_reward_zero_std": 0.5, "grad_norm": 0.35546875, "learning_rate": 9.773242630385487e-07, "loss": 0.0396, "num_tokens": 4258278.0, "reward": 0.1328125, "reward_std": 0.4386323094367981, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 41, "step_time": 290.2552304994315 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3830.0, "completions/mean_length": 1656.03125, "completions/mean_terminated_length": 1092.9615478515625, "completions/min_length": 139.0, "completions/min_terminated_length": 139.0, "entropy": 3.188207183033228, "epoch": 0.023809523809523808, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 9.767573696145124e-07, "loss": 0.0, "num_tokens": 4376928.0, "reward": 0.0, "reward_std": 0.0, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 42, "step_time": 301.859591527842 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4076.0, "completions/mean_length": 1559.265625, "completions/mean_terminated_length": 1032.7735595703125, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 3.5811468213796616, "epoch": 0.024376417233560092, "frac_reward_zero_std": 0.625, "grad_norm": 0.193359375, "learning_rate": 9.761904761904762e-07, "loss": 0.0355, "num_tokens": 4488233.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 43, "step_time": 290.75280567258596 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3616.0, "completions/mean_length": 1324.796875, "completions/mean_terminated_length": 928.9107666015625, "completions/min_length": 25.0, "completions/min_terminated_length": 25.0, "entropy": 3.2633891254663467, "epoch": 0.024943310657596373, "frac_reward_zero_std": 0.875, "grad_norm": 0.1640625, "learning_rate": 9.7562358276644e-07, "loss": 0.0085, "num_tokens": 4584436.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 44, "step_time": 292.3490004586056 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4083.0, "completions/mean_length": 1433.921875, "completions/mean_terminated_length": 940.9444580078125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 4.2754261419177055, "epoch": 0.025510204081632654, "frac_reward_zero_std": 0.75, "grad_norm": 0.2197265625, "learning_rate": 9.750566893424036e-07, "loss": 0.0036, "num_tokens": 4688903.0, "reward": 0.1015625, "reward_std": 0.42891231179237366, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 45, "step_time": 293.87602140940726 }, { "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": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3726.0, "completions/mean_length": 1855.015625, "completions/mean_terminated_length": 1227.5399169921875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 4.3639324232935905, "epoch": 0.026077097505668934, "frac_reward_zero_std": 0.75, "grad_norm": 0.1474609375, "learning_rate": 9.744897959183674e-07, "loss": 0.0168, "num_tokens": 4819584.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 46, "step_time": 294.2176110269502 }, { "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": 0.28125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3513.0, "completions/mean_length": 1799.515625, "completions/mean_terminated_length": 900.8912963867188, "completions/min_length": 163.0, "completions/min_terminated_length": 163.0, "entropy": 3.222260318696499, "epoch": 0.026643990929705215, "frac_reward_zero_std": 0.75, "grad_norm": 0.109375, "learning_rate": 9.739229024943311e-07, "loss": 0.007, "num_tokens": 4949033.0, "reward": 0.1015625, "reward_std": 0.40999117493629456, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 47, "step_time": 305.7597019765526 }, { "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": 4096.0, "completions/max_terminated_length": 4011.0, "completions/mean_length": 1215.359375, "completions/mean_terminated_length": 1023.3167114257812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.471595462411642, "epoch": 0.027210884353741496, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 9.733560090702947e-07, "loss": 0.0, "num_tokens": 5037408.0, "reward": 0.0, "reward_std": 0.0, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 48, "step_time": 289.0369003722444 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3974.0, "completions/mean_length": 1368.90625, "completions/mean_terminated_length": 979.3214721679688, "completions/min_length": 34.0, "completions/min_terminated_length": 34.0, "entropy": 4.266951911151409, "epoch": 0.027777777777777776, "frac_reward_zero_std": 0.875, "grad_norm": 0.18359375, "learning_rate": 9.727891156462584e-07, "loss": 0.0288, "num_tokens": 5135666.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 49, "step_time": 291.6630548238754 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3142.0, "completions/mean_length": 1372.0, "completions/mean_terminated_length": 743.3846435546875, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "entropy": 2.301131073385477, "epoch": 0.02834467120181406, "frac_reward_zero_std": 0.75, "grad_norm": 0.1650390625, "learning_rate": 9.722222222222222e-07, "loss": -0.0092, "num_tokens": 5235138.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 50, "step_time": 291.11338582262397 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3997.0, "completions/mean_length": 1138.609375, "completions/mean_terminated_length": 832.6724243164062, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "entropy": 2.9807814955711365, "epoch": 0.02891156462585034, "frac_reward_zero_std": 0.75, "grad_norm": 0.197265625, "learning_rate": 9.71655328798186e-07, "loss": 0.0139, "num_tokens": 5322953.0, "reward": 0.078125, "reward_std": 0.2847827076911926, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 51, "step_time": 336.76144039351493 }, { "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": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4023.0, "completions/mean_length": 1662.921875, "completions/mean_terminated_length": 1042.7255859375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.494113929569721, "epoch": 0.02947845804988662, "frac_reward_zero_std": 0.75, "grad_norm": 0.1298828125, "learning_rate": 9.710884353741497e-07, "loss": 0.0124, "num_tokens": 5440068.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 52, "step_time": 289.4927029591054 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3873.0, "completions/mean_length": 1544.671875, "completions/mean_terminated_length": 955.9038696289062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.09750372171402, "epoch": 0.030045351473922902, "frac_reward_zero_std": 0.75, "grad_norm": 0.2099609375, "learning_rate": 9.705215419501134e-07, "loss": -0.0162, "num_tokens": 5550303.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 53, "step_time": 291.7991081448272 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3854.0, "completions/mean_length": 1282.859375, "completions/mean_terminated_length": 937.385986328125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.4761949479579926, "epoch": 0.030612244897959183, "frac_reward_zero_std": 0.75, "grad_norm": 0.1748046875, "learning_rate": 9.69954648526077e-07, "loss": -0.0001, "num_tokens": 5644734.0, "reward": 0.2421875, "reward_std": 0.7126531600952148, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 54, "step_time": 293.1030203923583 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3929.0, "completions/mean_length": 1217.65625, "completions/mean_terminated_length": 864.1754150390625, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 2.805364416912198, "epoch": 0.031179138321995464, "frac_reward_zero_std": 0.875, "grad_norm": 0.1591796875, "learning_rate": 9.693877551020407e-07, "loss": 0.0099, "num_tokens": 5735168.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 55, "step_time": 292.36450640298426 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3811.0, "completions/mean_length": 1605.859375, "completions/mean_terminated_length": 1144.7222900390625, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "entropy": 3.727115612477064, "epoch": 0.031746031746031744, "frac_reward_zero_std": 0.875, "grad_norm": 0.125, "learning_rate": 9.688208616780045e-07, "loss": 0.0173, "num_tokens": 5850943.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 56, "step_time": 293.6549425115809 }, { "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": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3305.0, "completions/mean_length": 1580.921875, "completions/mean_terminated_length": 939.8235473632812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.698149297386408, "epoch": 0.03231292517006803, "frac_reward_zero_std": 0.875, "grad_norm": 0.14453125, "learning_rate": 9.682539682539682e-07, "loss": 0.044, "num_tokens": 5966306.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 57, "step_time": 305.46879615914077 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3931.0, "completions/mean_length": 1036.453125, "completions/mean_terminated_length": 719.9483032226562, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.0936585515737534, "epoch": 0.032879818594104306, "frac_reward_zero_std": 0.5, "grad_norm": 0.3359375, "learning_rate": 9.67687074829932e-07, "loss": -0.0214, "num_tokens": 6043111.0, "reward": 0.2578125, "reward_std": 0.629891574382782, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 58, "step_time": 288.52694783732295 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3640.0, "completions/mean_length": 1435.484375, "completions/mean_terminated_length": 1108.75439453125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.869808528572321, "epoch": 0.03344671201814059, "frac_reward_zero_std": 0.875, "grad_norm": 0.169921875, "learning_rate": 9.671201814058957e-07, "loss": 0.0258, "num_tokens": 6146982.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 59, "step_time": 292.3406807044521 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3907.0, "completions/mean_length": 1457.140625, "completions/mean_terminated_length": 1080.1607666015625, "completions/min_length": 100.0, "completions/min_terminated_length": 100.0, "entropy": 4.028498370200396, "epoch": 0.034013605442176874, "frac_reward_zero_std": 0.75, "grad_norm": 0.1796875, "learning_rate": 9.665532879818595e-07, "loss": 0.0276, "num_tokens": 6252479.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 60, "step_time": 295.60633749328554 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4054.0, "completions/mean_length": 1553.203125, "completions/mean_terminated_length": 1082.3148193359375, "completions/min_length": 195.0, "completions/min_terminated_length": 195.0, "entropy": 3.7958782464265823, "epoch": 0.03458049886621315, "frac_reward_zero_std": 0.875, "grad_norm": 0.1767578125, "learning_rate": 9.65986394557823e-07, "loss": -0.004, "num_tokens": 6364244.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 61, "step_time": 299.2967924810946 }, { "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": 0.265625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3916.0, "completions/mean_length": 1934.71875, "completions/mean_terminated_length": 1152.9786376953125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 5.334939103573561, "epoch": 0.035147392290249435, "frac_reward_zero_std": 0.75, "grad_norm": 0.1083984375, "learning_rate": 9.654195011337868e-07, "loss": -0.0012, "num_tokens": 6498050.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 62, "step_time": 291.10220227204263 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4055.0, "completions/mean_length": 1416.6875, "completions/mean_terminated_length": 920.5185546875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.310685142874718, "epoch": 0.03571428571428571, "frac_reward_zero_std": 0.875, "grad_norm": 0.11767578125, "learning_rate": 9.648526077097505e-07, "loss": 0.0242, "num_tokens": 6599990.0, "reward": 0.0625, "reward_std": 0.35073620080947876, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 63, "step_time": 290.805386999622 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4052.0, "completions/mean_length": 1795.5, "completions/mean_terminated_length": 1264.615478515625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.7707428578287363, "epoch": 0.036281179138321996, "frac_reward_zero_std": 0.875, "grad_norm": 0.11279296875, "learning_rate": 9.642857142857142e-07, "loss": 0.0101, "num_tokens": 6728598.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 64, "step_time": 297.35328650381416 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4065.0, "completions/mean_length": 1198.28125, "completions/mean_terminated_length": 784.3214721679688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.7539037615060806, "epoch": 0.036848072562358274, "frac_reward_zero_std": 0.625, "grad_norm": 0.2373046875, "learning_rate": 9.63718820861678e-07, "loss": 0.0287, "num_tokens": 6818008.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 65, "step_time": 294.24104030337185 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3850.0, "completions/mean_length": 1337.84375, "completions/mean_terminated_length": 1052.5172119140625, "completions/min_length": 55.0, "completions/min_terminated_length": 55.0, "entropy": 3.438830465078354, "epoch": 0.03741496598639456, "frac_reward_zero_std": 0.875, "grad_norm": 0.126953125, "learning_rate": 9.631519274376417e-07, "loss": -0.0048, "num_tokens": 6917318.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 66, "step_time": 296.2423573601991 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3670.0, "completions/mean_length": 1163.953125, "completions/mean_terminated_length": 915.4745483398438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.012999653816223, "epoch": 0.03798185941043084, "frac_reward_zero_std": 0.5, "grad_norm": 0.32421875, "learning_rate": 9.625850340136053e-07, "loss": 0.0592, "num_tokens": 7003571.0, "reward": 0.140625, "reward_std": 0.4917473793029785, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 67, "step_time": 291.4204217866063 }, { "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": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3203.0, "completions/mean_length": 1563.03125, "completions/mean_terminated_length": 917.37255859375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.199941698461771, "epoch": 0.03854875283446712, "frac_reward_zero_std": 0.75, "grad_norm": 0.1875, "learning_rate": 9.62018140589569e-07, "loss": -0.0126, "num_tokens": 7115117.0, "reward": 0.1171875, "reward_std": 0.43408530950546265, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 68, "step_time": 292.0785713447258 }, { "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": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 2907.0, "completions/mean_length": 1363.328125, "completions/mean_terminated_length": 598.1799926757812, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 1.8701567724347115, "epoch": 0.0391156462585034, "frac_reward_zero_std": 0.625, "grad_norm": 0.193359375, "learning_rate": 9.614512471655328e-07, "loss": 0.0002, "num_tokens": 7215274.0, "reward": 0.1640625, "reward_std": 0.496453195810318, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 69, "step_time": 292.96778441406786 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4004.0, "completions/mean_length": 1443.53125, "completions/mean_terminated_length": 952.3333129882812, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "entropy": 3.6688392758369446, "epoch": 0.03968253968253968, "frac_reward_zero_std": 0.875, "grad_norm": 0.10498046875, "learning_rate": 9.608843537414965e-07, "loss": 0.0001, "num_tokens": 7319268.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 70, "step_time": 289.6370008327067 }, { "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": 4096.0, "completions/max_terminated_length": 3507.0, "completions/mean_length": 955.609375, "completions/mean_terminated_length": 746.2500610351562, "completions/min_length": 140.0, "completions/min_terminated_length": 140.0, "entropy": 3.290744874626398, "epoch": 0.040249433106575964, "frac_reward_zero_std": 0.625, "grad_norm": 0.373046875, "learning_rate": 9.603174603174603e-07, "loss": -0.053, "num_tokens": 7391395.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 71, "step_time": 289.79417512752116 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3746.0, "completions/mean_length": 996.6875, "completions/mean_terminated_length": 844.2622680664062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.150678135454655, "epoch": 0.04081632653061224, "frac_reward_zero_std": 0.75, "grad_norm": 0.2490234375, "learning_rate": 9.59750566893424e-07, "loss": 0.0268, "num_tokens": 7464487.0, "reward": 0.1015625, "reward_std": 0.42891231179237366, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 72, "step_time": 285.74950027186424 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3726.0, "completions/mean_length": 1227.15625, "completions/mean_terminated_length": 930.3793334960938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.171508941799402, "epoch": 0.041383219954648526, "frac_reward_zero_std": 0.5, "grad_norm": 0.30078125, "learning_rate": 9.591836734693876e-07, "loss": -0.0018, "num_tokens": 7562281.0, "reward": 0.171875, "reward_std": 0.49776285886764526, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 73, "step_time": 370.6855717310682 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3777.0, "completions/mean_length": 1035.265625, "completions/mean_terminated_length": 659.385986328125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.716996032744646, "epoch": 0.04195011337868481, "frac_reward_zero_std": 0.625, "grad_norm": 0.3671875, "learning_rate": 9.586167800453513e-07, "loss": -0.073, "num_tokens": 7639194.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 74, "step_time": 287.5007770322263 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3763.0, "completions/mean_length": 1367.421875, "completions/mean_terminated_length": 862.129638671875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.463905990123749, "epoch": 0.04251700680272109, "frac_reward_zero_std": 0.625, "grad_norm": 0.2333984375, "learning_rate": 9.58049886621315e-07, "loss": 0.009, "num_tokens": 7738213.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 75, "step_time": 292.26083657797426 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3623.0, "completions/mean_length": 908.25, "completions/mean_terminated_length": 638.1016845703125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.7447003312408924, "epoch": 0.04308390022675737, "frac_reward_zero_std": 0.5, "grad_norm": 0.33203125, "learning_rate": 9.574829931972788e-07, "loss": 0.0071, "num_tokens": 7806445.0, "reward": 0.328125, "reward_std": 0.7303985953330994, "rewards/accuracy_reward_func/mean": 0.3125, "rewards/accuracy_reward_func/std": 0.7319250702857971, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 76, "step_time": 286.0780157428235 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4011.0, "completions/mean_length": 1678.390625, "completions/mean_terminated_length": 1230.6851806640625, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "entropy": 5.76315838098526, "epoch": 0.04365079365079365, "frac_reward_zero_std": 0.75, "grad_norm": 0.2041015625, "learning_rate": 9.569160997732426e-07, "loss": 0.0146, "num_tokens": 7925038.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 77, "step_time": 289.65414382983 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3810.0, "completions/mean_length": 1468.625, "completions/mean_terminated_length": 1038.69091796875, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 2.9119240194559097, "epoch": 0.04421768707482993, "frac_reward_zero_std": 0.75, "grad_norm": 0.2119140625, "learning_rate": 9.563492063492063e-07, "loss": 0.0445, "num_tokens": 8031758.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 78, "step_time": 295.3073107143864 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3986.0, "completions/mean_length": 1495.8125, "completions/mean_terminated_length": 895.769287109375, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "entropy": 2.7026463598012924, "epoch": 0.04478458049886622, "frac_reward_zero_std": 0.875, "grad_norm": 0.1123046875, "learning_rate": 9.5578231292517e-07, "loss": -0.0189, "num_tokens": 8139346.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 79, "step_time": 292.16124310158193 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3691.0, "completions/mean_length": 1399.546875, "completions/mean_terminated_length": 958.30908203125, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "entropy": 3.1997539959847927, "epoch": 0.045351473922902494, "frac_reward_zero_std": 0.75, "grad_norm": 0.1826171875, "learning_rate": 9.552154195011338e-07, "loss": 0.0371, "num_tokens": 8240445.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 80, "step_time": 289.87046337500215 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3773.0, "completions/mean_length": 1490.546875, "completions/mean_terminated_length": 1008.0555419921875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.017605835571885, "epoch": 0.04591836734693878, "frac_reward_zero_std": 0.5, "grad_norm": 0.248046875, "learning_rate": 9.546485260770974e-07, "loss": 0.0258, "num_tokens": 8346640.0, "reward": 0.1328125, "reward_std": 0.489915668964386, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 81, "step_time": 288.093112712726 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3936.0, "completions/mean_length": 1136.359375, "completions/mean_terminated_length": 885.5423583984375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.24729186668992, "epoch": 0.046485260770975055, "frac_reward_zero_std": 0.875, "grad_norm": 0.201171875, "learning_rate": 9.54081632653061e-07, "loss": 0.0146, "num_tokens": 8435103.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 82, "step_time": 324.48642648383975 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3430.0, "completions/mean_length": 1098.140625, "completions/mean_terminated_length": 729.9824829101562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.5958854481577873, "epoch": 0.04705215419501134, "frac_reward_zero_std": 0.625, "grad_norm": 0.306640625, "learning_rate": 9.53514739229025e-07, "loss": -0.0005, "num_tokens": 8515712.0, "reward": 0.1171875, "reward_std": 0.43408530950546265, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 83, "step_time": 289.2759247524664 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3964.0, "completions/mean_length": 1243.140625, "completions/mean_terminated_length": 948.0172119140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.397062368690968, "epoch": 0.047619047619047616, "frac_reward_zero_std": 0.875, "grad_norm": 0.1904296875, "learning_rate": 9.529478458049887e-07, "loss": -0.0087, "num_tokens": 8607217.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 84, "step_time": 291.1187508981675 }, { "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": 0.234375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4079.0, "completions/mean_length": 1913.15625, "completions/mean_terminated_length": 1244.938720703125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.4104487486183643, "epoch": 0.0481859410430839, "frac_reward_zero_std": 0.875, "grad_norm": 0.10302734375, "learning_rate": 9.523809523809522e-07, "loss": -0.0167, "num_tokens": 8742627.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 85, "step_time": 291.9923343649134 }, { "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": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3797.0, "completions/mean_length": 1716.296875, "completions/mean_terminated_length": 1049.97998046875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.2259551621973515, "epoch": 0.048752834467120185, "frac_reward_zero_std": 0.875, "grad_norm": 0.1025390625, "learning_rate": 9.51814058956916e-07, "loss": -0.0127, "num_tokens": 8864598.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 86, "step_time": 293.8602614365518 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3875.0, "completions/mean_length": 1328.734375, "completions/mean_terminated_length": 933.4107666015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.2866496331989765, "epoch": 0.04931972789115646, "frac_reward_zero_std": 0.875, "grad_norm": 0.12451171875, "learning_rate": 9.512471655328797e-07, "loss": 0.0046, "num_tokens": 8963653.0, "reward": 0.125, "reward_std": 0.48795002698898315, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 87, "step_time": 298.11381549574435 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4076.0, "completions/mean_length": 1145.359375, "completions/mean_terminated_length": 783.0, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.2719305753707886, "epoch": 0.049886621315192746, "frac_reward_zero_std": 0.875, "grad_norm": 0.18359375, "learning_rate": 9.506802721088435e-07, "loss": -0.0021, "num_tokens": 9057756.0, "reward": 0.046875, "reward_std": 0.14689241349697113, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 88, "step_time": 366.73505029734224 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3767.0, "completions/mean_length": 1474.8125, "completions/mean_terminated_length": 989.4074096679688, "completions/min_length": 122.0, "completions/min_terminated_length": 122.0, "entropy": 3.3253636807203293, "epoch": 0.05045351473922902, "frac_reward_zero_std": 0.875, "grad_norm": 0.16796875, "learning_rate": 9.501133786848072e-07, "loss": 0.0261, "num_tokens": 9164160.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 89, "step_time": 297.5648354878649 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4073.0, "completions/mean_length": 1280.140625, "completions/mean_terminated_length": 758.6851806640625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.639719892293215, "epoch": 0.05102040816326531, "frac_reward_zero_std": 0.625, "grad_norm": 0.26171875, "learning_rate": 9.49546485260771e-07, "loss": 0.017, "num_tokens": 9258785.0, "reward": 0.2109375, "reward_std": 0.5896238088607788, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 90, "step_time": 292.0028676353395 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3221.0, "completions/mean_length": 1026.84375, "completions/mean_terminated_length": 709.3448486328125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.548398684710264, "epoch": 0.051587301587301584, "frac_reward_zero_std": 0.75, "grad_norm": 0.2255859375, "learning_rate": 9.489795918367347e-07, "loss": -0.0125, "num_tokens": 9338359.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 91, "step_time": 306.1025436343625 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3879.0, "completions/mean_length": 1430.9375, "completions/mean_terminated_length": 877.8113403320312, "completions/min_length": 65.0, "completions/min_terminated_length": 65.0, "entropy": 3.525468062609434, "epoch": 0.05215419501133787, "frac_reward_zero_std": 0.75, "grad_norm": 0.1689453125, "learning_rate": 9.484126984126984e-07, "loss": 0.0098, "num_tokens": 9441339.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 92, "step_time": 292.2532150428742 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4031.0, "completions/mean_length": 1503.125, "completions/mean_terminated_length": 964.9811401367188, "completions/min_length": 106.0, "completions/min_terminated_length": 106.0, "entropy": 3.780000723898411, "epoch": 0.05272108843537415, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 9.47845804988662e-07, "loss": 0.0, "num_tokens": 9550755.0, "reward": 0.0, "reward_std": 0.0, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 93, "step_time": 301.9815172245726 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4078.0, "completions/mean_length": 1439.8125, "completions/mean_terminated_length": 888.5283203125, "completions/min_length": 27.0, "completions/min_terminated_length": 27.0, "entropy": 4.099758766591549, "epoch": 0.05328798185941043, "frac_reward_zero_std": 0.875, "grad_norm": 0.162109375, "learning_rate": 9.472789115646258e-07, "loss": 0.0362, "num_tokens": 9655567.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 94, "step_time": 295.015036133118 }, { "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": 4096.0, "completions/max_terminated_length": 4018.0, "completions/mean_length": 1064.84375, "completions/mean_terminated_length": 862.7667236328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.8394701406359673, "epoch": 0.053854875283446714, "frac_reward_zero_std": 0.75, "grad_norm": 0.287109375, "learning_rate": 9.467120181405895e-07, "loss": 0.0152, "num_tokens": 9734885.0, "reward": 0.078125, "reward_std": 0.4005824029445648, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 95, "step_time": 290.8204828929156 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3964.0, "completions/mean_length": 1292.1875, "completions/mean_terminated_length": 772.9629516601562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.363392487168312, "epoch": 0.05442176870748299, "frac_reward_zero_std": 0.75, "grad_norm": 0.21484375, "learning_rate": 9.461451247165533e-07, "loss": 0.0659, "num_tokens": 9829737.0, "reward": 0.140625, "reward_std": 0.4753340482711792, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 96, "step_time": 298.31369636021554 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3902.0, "completions/mean_length": 1551.15625, "completions/mean_terminated_length": 1079.888916015625, "completions/min_length": 183.0, "completions/min_terminated_length": 183.0, "entropy": 3.6113050170242786, "epoch": 0.054988662131519275, "frac_reward_zero_std": 0.875, "grad_norm": 0.1689453125, "learning_rate": 9.45578231292517e-07, "loss": 0.0202, "num_tokens": 9941531.0, "reward": 0.1015625, "reward_std": 0.46444785594940186, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 97, "step_time": 296.1742769367993 }, { "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": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3885.0, "completions/mean_length": 1697.015625, "completions/mean_terminated_length": 1025.2999267578125, "completions/min_length": 203.0, "completions/min_terminated_length": 203.0, "entropy": 3.0438320823013783, "epoch": 0.05555555555555555, "frac_reward_zero_std": 0.875, "grad_norm": 0.1220703125, "learning_rate": 9.450113378684807e-07, "loss": -0.0024, "num_tokens": 10062308.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 98, "step_time": 291.39408755209297 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3155.0, "completions/mean_length": 1273.40625, "completions/mean_terminated_length": 811.5272216796875, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 3.6760238520801067, "epoch": 0.05612244897959184, "frac_reward_zero_std": 0.75, "grad_norm": 0.216796875, "learning_rate": 9.444444444444444e-07, "loss": 0.0585, "num_tokens": 10158086.0, "reward": 0.1484375, "reward_std": 0.49344661831855774, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 99, "step_time": 317.7352059157565 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3749.0, "completions/mean_length": 1897.875, "completions/mean_terminated_length": 1441.660400390625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 5.14299500733614, "epoch": 0.05668934240362812, "frac_reward_zero_std": 0.875, "grad_norm": 0.126953125, "learning_rate": 9.438775510204082e-07, "loss": 0.016, "num_tokens": 10291630.0, "reward": 0.09375, "reward_std": 0.4443204402923584, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 100, "step_time": 294.5599527070299 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3817.0, "completions/mean_length": 1232.4375, "completions/mean_terminated_length": 763.8544921875, "completions/min_length": 29.0, "completions/min_terminated_length": 29.0, "entropy": 3.398377440869808, "epoch": 0.0572562358276644, "frac_reward_zero_std": 0.75, "grad_norm": 0.220703125, "learning_rate": 9.433106575963718e-07, "loss": 0.0437, "num_tokens": 10381466.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 101, "step_time": 290.80642426013947 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3059.0, "completions/mean_length": 1041.484375, "completions/mean_terminated_length": 725.5, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "entropy": 2.0532282926142216, "epoch": 0.05782312925170068, "frac_reward_zero_std": 0.625, "grad_norm": 0.251953125, "learning_rate": 9.427437641723356e-07, "loss": -0.0042, "num_tokens": 10459361.0, "reward": 0.125, "reward_std": 0.4364357888698578, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 102, "step_time": 288.7475468451157 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3790.0, "completions/mean_length": 1440.09375, "completions/mean_terminated_length": 1005.4909057617188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.355026971548796, "epoch": 0.05839002267573696, "frac_reward_zero_std": 0.875, "grad_norm": 0.15234375, "learning_rate": 9.421768707482993e-07, "loss": 0.0284, "num_tokens": 10562695.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 103, "step_time": 290.79865114577115 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3883.0, "completions/mean_length": 1285.4375, "completions/mean_terminated_length": 825.5272216796875, "completions/min_length": 19.0, "completions/min_terminated_length": 19.0, "entropy": 3.470572918653488, "epoch": 0.05895691609977324, "frac_reward_zero_std": 0.875, "grad_norm": 0.1103515625, "learning_rate": 9.41609977324263e-07, "loss": 0.0003, "num_tokens": 10658595.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 104, "step_time": 298.1190188927576 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4069.0, "completions/mean_length": 1428.265625, "completions/mean_terminated_length": 874.5848999023438, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 3.4206804297864437, "epoch": 0.05952380952380952, "frac_reward_zero_std": 0.75, "grad_norm": 0.1689453125, "learning_rate": 9.410430839002267e-07, "loss": 0.0263, "num_tokens": 10760788.0, "reward": 0.125, "reward_std": 0.48795002698898315, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 105, "step_time": 290.42720403149724 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3881.0, "completions/mean_length": 1183.28125, "completions/mean_terminated_length": 936.440673828125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.111788738518953, "epoch": 0.060090702947845805, "frac_reward_zero_std": 0.625, "grad_norm": 0.3515625, "learning_rate": 9.404761904761904e-07, "loss": -0.0509, "num_tokens": 10847494.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 106, "step_time": 291.5891627203673 }, { "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": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3815.0, "completions/mean_length": 1559.203125, "completions/mean_terminated_length": 912.5686645507812, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 2.9711022675037384, "epoch": 0.06065759637188209, "frac_reward_zero_std": 0.75, "grad_norm": 0.1494140625, "learning_rate": 9.399092970521542e-07, "loss": -0.0135, "num_tokens": 10960571.0, "reward": 0.2265625, "reward_std": 0.6419816017150879, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 107, "step_time": 299.18715846166015 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3181.0, "completions/mean_length": 1501.265625, "completions/mean_terminated_length": 1020.75927734375, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "entropy": 3.357049345970154, "epoch": 0.061224489795918366, "frac_reward_zero_std": 0.875, "grad_norm": 0.1640625, "learning_rate": 9.393424036281179e-07, "loss": -0.0042, "num_tokens": 11067268.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 108, "step_time": 290.50813248194754 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3604.0, "completions/mean_length": 967.21875, "completions/mean_terminated_length": 643.5516967773438, "completions/min_length": 23.0, "completions/min_terminated_length": 23.0, "entropy": 3.078561846166849, "epoch": 0.06179138321995465, "frac_reward_zero_std": 0.625, "grad_norm": 0.287109375, "learning_rate": 9.387755102040816e-07, "loss": 0.0282, "num_tokens": 11139610.0, "reward": 0.1015625, "reward_std": 0.42891231179237366, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 109, "step_time": 290.01873538829386 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4024.0, "completions/mean_length": 1146.640625, "completions/mean_terminated_length": 784.4385986328125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.0737808886915445, "epoch": 0.06235827664399093, "frac_reward_zero_std": 0.75, "grad_norm": 0.1845703125, "learning_rate": 9.382086167800452e-07, "loss": 0.0355, "num_tokens": 11224219.0, "reward": 0.1015625, "reward_std": 0.40999117493629456, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 110, "step_time": 289.6260637482628 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3743.0, "completions/mean_length": 1310.640625, "completions/mean_terminated_length": 854.8544921875, "completions/min_length": 21.0, "completions/min_terminated_length": 21.0, "entropy": 3.687939263880253, "epoch": 0.06292517006802721, "frac_reward_zero_std": 0.875, "grad_norm": 0.162109375, "learning_rate": 9.37641723356009e-07, "loss": 0.0378, "num_tokens": 11321404.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 111, "step_time": 302.7629135893658 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 1904.0, "completions/mean_length": 1221.203125, "completions/mean_terminated_length": 557.7884521484375, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 1.8763416595757008, "epoch": 0.06349206349206349, "frac_reward_zero_std": 0.875, "grad_norm": 0.10595703125, "learning_rate": 9.370748299319727e-07, "loss": 0.0049, "num_tokens": 11411185.0, "reward": 0.109375, "reward_std": 0.4495919644832611, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 112, "step_time": 293.1259895954281 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3994.0, "completions/mean_length": 1544.921875, "completions/mean_terminated_length": 956.2115478515625, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "entropy": 4.175659708678722, "epoch": 0.06405895691609978, "frac_reward_zero_std": 0.875, "grad_norm": 0.07763671875, "learning_rate": 9.365079365079365e-07, "loss": -0.0035, "num_tokens": 11522804.0, "reward": 0.140625, "reward_std": 0.5379911661148071, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 113, "step_time": 308.48071416281164 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3717.0, "completions/mean_length": 1413.578125, "completions/mean_terminated_length": 856.8490600585938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.534675944596529, "epoch": 0.06462585034013606, "frac_reward_zero_std": 0.625, "grad_norm": 0.298828125, "learning_rate": 9.359410430839002e-07, "loss": -0.0452, "num_tokens": 11626985.0, "reward": 0.1640625, "reward_std": 0.496453195810318, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 114, "step_time": 309.24449787382036 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3788.0, "completions/mean_length": 1072.546875, "completions/mean_terminated_length": 759.77587890625, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.437642801553011, "epoch": 0.06519274376417233, "frac_reward_zero_std": 0.375, "grad_norm": 0.40234375, "learning_rate": 9.35374149659864e-07, "loss": -0.0241, "num_tokens": 11707180.0, "reward": 0.390625, "reward_std": 0.7841117978096008, "rewards/accuracy_reward_func/mean": 0.375, "rewards/accuracy_reward_func/std": 0.7867957949638367, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 115, "step_time": 293.03551427833736 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3891.0, "completions/mean_length": 950.890625, "completions/mean_terminated_length": 684.35595703125, "completions/min_length": 120.0, "completions/min_terminated_length": 120.0, "entropy": 2.9594995956867933, "epoch": 0.06575963718820861, "frac_reward_zero_std": 0.5, "grad_norm": 0.359375, "learning_rate": 9.348072562358276e-07, "loss": 0.0498, "num_tokens": 11785261.0, "reward": 0.1875, "reward_std": 0.5455447435379028, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 116, "step_time": 334.32934977766126 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4081.0, "completions/mean_length": 1303.171875, "completions/mean_terminated_length": 785.9815063476562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.3868199810385704, "epoch": 0.0663265306122449, "frac_reward_zero_std": 0.875, "grad_norm": 0.1630859375, "learning_rate": 9.342403628117913e-07, "loss": 0.0085, "num_tokens": 11882896.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 117, "step_time": 309.05003222823143 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4043.0, "completions/mean_length": 1259.1875, "completions/mean_terminated_length": 794.9818115234375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.722678354009986, "epoch": 0.06689342403628118, "frac_reward_zero_std": 0.875, "grad_norm": 0.1015625, "learning_rate": 9.33673469387755e-07, "loss": -0.0001, "num_tokens": 11976516.0, "reward": 0.09375, "reward_std": 0.4443204402923584, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 118, "step_time": 306.37148281373084 }, { "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": 0.03125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3668.0, "completions/mean_length": 917.4375, "completions/mean_terminated_length": 814.9031982421875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.591132014989853, "epoch": 0.06746031746031746, "frac_reward_zero_std": 0.75, "grad_norm": 0.333984375, "learning_rate": 9.331065759637188e-07, "loss": 0.0081, "num_tokens": 12045200.0, "reward": 0.125, "reward_std": 0.4364357888698578, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 119, "step_time": 287.9365804567933 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3345.0, "completions/mean_length": 1416.71875, "completions/mean_terminated_length": 978.2908935546875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.8865617848932743, "epoch": 0.06802721088435375, "frac_reward_zero_std": 0.875, "grad_norm": 0.1337890625, "learning_rate": 9.325396825396825e-07, "loss": 0.0316, "num_tokens": 12147654.0, "reward": 0.0625, "reward_std": 0.35073620080947876, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 120, "step_time": 291.8641989789903 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3760.0, "completions/mean_length": 1235.0625, "completions/mean_terminated_length": 826.357177734375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.7196299508213997, "epoch": 0.06859410430839002, "frac_reward_zero_std": 0.875, "grad_norm": 0.177734375, "learning_rate": 9.319727891156463e-07, "loss": 0.0283, "num_tokens": 12239306.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 121, "step_time": 297.73565461300313 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3821.0, "completions/mean_length": 1245.46875, "completions/mean_terminated_length": 838.2500610351562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.7032400891184807, "epoch": 0.0691609977324263, "frac_reward_zero_std": 0.875, "grad_norm": 0.1279296875, "learning_rate": 9.3140589569161e-07, "loss": 0.0156, "num_tokens": 12331224.0, "reward": 0.09375, "reward_std": 0.426084041595459, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 122, "step_time": 296.430201029405 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3816.0, "completions/mean_length": 1230.9375, "completions/mean_terminated_length": 821.6428833007812, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 3.972167894244194, "epoch": 0.06972789115646258, "frac_reward_zero_std": 0.625, "grad_norm": 0.2099609375, "learning_rate": 9.308390022675737e-07, "loss": 0.0124, "num_tokens": 12424828.0, "reward": 0.203125, "reward_std": 0.5471335649490356, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 123, "step_time": 303.4636345040053 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3748.0, "completions/mean_length": 1233.9375, "completions/mean_terminated_length": 937.862060546875, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "entropy": 5.474953629076481, "epoch": 0.07029478458049887, "frac_reward_zero_std": 0.75, "grad_norm": 0.1982421875, "learning_rate": 9.302721088435374e-07, "loss": 0.0386, "num_tokens": 12514768.0, "reward": 0.109375, "reward_std": 0.4315783977508545, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 124, "step_time": 291.9685418345034 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3993.0, "completions/mean_length": 1272.046875, "completions/mean_terminated_length": 749.0925903320312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.9104819521307945, "epoch": 0.07086167800453515, "frac_reward_zero_std": 0.875, "grad_norm": 0.126953125, "learning_rate": 9.29705215419501e-07, "loss": 0.0005, "num_tokens": 12609787.0, "reward": 0.0546875, "reward_std": 0.15728822350502014, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 125, "step_time": 297.0362865040079 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4053.0, "completions/mean_length": 1407.515625, "completions/mean_terminated_length": 1023.4464721679688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.175208568572998, "epoch": 0.07142857142857142, "frac_reward_zero_std": 0.875, "grad_norm": 0.1494140625, "learning_rate": 9.291383219954648e-07, "loss": -0.0045, "num_tokens": 12711820.0, "reward": 0.0625, "reward_std": 0.32732683420181274, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 126, "step_time": 295.77356148231775 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3870.0, "completions/mean_length": 1636.09375, "completions/mean_terminated_length": 1068.423095703125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.013583540916443, "epoch": 0.07199546485260772, "frac_reward_zero_std": 0.625, "grad_norm": 0.2138671875, "learning_rate": 9.285714285714285e-07, "loss": 0.0119, "num_tokens": 12828914.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 127, "step_time": 292.88032844942063 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3835.0, "completions/mean_length": 1292.484375, "completions/mean_terminated_length": 833.7272338867188, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.8224235624074936, "epoch": 0.07256235827664399, "frac_reward_zero_std": 0.875, "grad_norm": 0.16015625, "learning_rate": 9.280045351473923e-07, "loss": -0.0077, "num_tokens": 12923441.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 128, "step_time": 292.5007902001962 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3505.0, "completions/mean_length": 1495.3125, "completions/mean_terminated_length": 1069.745361328125, "completions/min_length": 123.0, "completions/min_terminated_length": 123.0, "entropy": 3.852652546018362, "epoch": 0.07312925170068027, "frac_reward_zero_std": 0.5, "grad_norm": 0.26171875, "learning_rate": 9.274376417233559e-07, "loss": 0.0085, "num_tokens": 13030005.0, "reward": 0.1953125, "reward_std": 0.5883605480194092, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 129, "step_time": 291.0363548276946 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3946.0, "completions/mean_length": 1184.90625, "completions/mean_terminated_length": 827.4035034179688, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 3.47156435623765, "epoch": 0.07369614512471655, "frac_reward_zero_std": 0.75, "grad_norm": 0.1689453125, "learning_rate": 9.268707482993197e-07, "loss": 0.0058, "num_tokens": 13117351.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 130, "step_time": 292.4133914411068 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3938.0, "completions/mean_length": 1478.46875, "completions/mean_terminated_length": 874.423095703125, "completions/min_length": 106.0, "completions/min_terminated_length": 106.0, "entropy": 4.215153083205223, "epoch": 0.07426303854875284, "frac_reward_zero_std": 0.75, "grad_norm": 0.1513671875, "learning_rate": 9.263038548752834e-07, "loss": 0.052, "num_tokens": 13223901.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 131, "step_time": 292.50151511095464 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3928.0, "completions/mean_length": 1487.890625, "completions/mean_terminated_length": 886.019287109375, "completions/min_length": 196.0, "completions/min_terminated_length": 196.0, "entropy": 2.6708176881074905, "epoch": 0.07482993197278912, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 9.257369614512472e-07, "loss": 0.0, "num_tokens": 13330694.0, "reward": 0.0625, "reward_std": 0.1666666716337204, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 132, "step_time": 298.96516027487814 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3714.0, "completions/mean_length": 1172.875, "completions/mean_terminated_length": 755.2857666015625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.068033292889595, "epoch": 0.07539682539682539, "frac_reward_zero_std": 0.75, "grad_norm": 0.255859375, "learning_rate": 9.251700680272108e-07, "loss": 0.0296, "num_tokens": 13417414.0, "reward": 0.140625, "reward_std": 0.4406772553920746, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 133, "step_time": 293.01860427390784 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3240.0, "completions/mean_length": 1434.46875, "completions/mean_terminated_length": 882.0755004882812, "completions/min_length": 107.0, "completions/min_terminated_length": 107.0, "entropy": 3.6458799690008163, "epoch": 0.07596371882086168, "frac_reward_zero_std": 0.75, "grad_norm": 0.197265625, "learning_rate": 9.246031746031746e-07, "loss": 0.0087, "num_tokens": 13521444.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 134, "step_time": 294.1661543697119 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3631.0, "completions/mean_length": 1475.046875, "completions/mean_terminated_length": 931.0755004882812, "completions/min_length": 186.0, "completions/min_terminated_length": 186.0, "entropy": 3.8298275284469128, "epoch": 0.07653061224489796, "frac_reward_zero_std": 0.5, "grad_norm": 0.2255859375, "learning_rate": 9.240362811791382e-07, "loss": 0.0435, "num_tokens": 13626319.0, "reward": 0.2109375, "reward_std": 0.5896238088607788, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 135, "step_time": 289.27914577163756 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3664.0, "completions/mean_length": 1512.859375, "completions/mean_terminated_length": 976.73583984375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.6895174458622932, "epoch": 0.07709750566893424, "frac_reward_zero_std": 0.75, "grad_norm": 0.18359375, "learning_rate": 9.23469387755102e-07, "loss": -0.0153, "num_tokens": 13734222.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 136, "step_time": 289.8237189883366 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3669.0, "completions/mean_length": 1097.0, "completions/mean_terminated_length": 606.2545166015625, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 2.4512119479477406, "epoch": 0.07766439909297052, "frac_reward_zero_std": 0.625, "grad_norm": 0.265625, "learning_rate": 9.229024943310657e-07, "loss": 0.0099, "num_tokens": 13816350.0, "reward": 0.296875, "reward_std": 0.7331100702285767, "rewards/accuracy_reward_func/mean": 0.25, "rewards/accuracy_reward_func/std": 0.6666666865348816, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 137, "step_time": 291.9875819766894 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4032.0, "completions/mean_length": 1222.3125, "completions/mean_terminated_length": 925.0344848632812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.987869266420603, "epoch": 0.0782312925170068, "frac_reward_zero_std": 0.5, "grad_norm": 0.33203125, "learning_rate": 9.223356009070295e-07, "loss": 0.0276, "num_tokens": 13909026.0, "reward": 0.3046875, "reward_std": 0.6993034482002258, "rewards/accuracy_reward_func/mean": 0.28125, "rewards/accuracy_reward_func/std": 0.7007648944854736, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 138, "step_time": 312.7329962141812 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3912.0, "completions/mean_length": 1212.65625, "completions/mean_terminated_length": 740.8363647460938, "completions/min_length": 89.0, "completions/min_terminated_length": 89.0, "entropy": 3.06102542206645, "epoch": 0.07879818594104308, "frac_reward_zero_std": 0.625, "grad_norm": 0.267578125, "learning_rate": 9.217687074829932e-07, "loss": 0.0942, "num_tokens": 13999660.0, "reward": 0.125, "reward_std": 0.4364357888698578, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 139, "step_time": 293.98899002652615 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3959.0, "completions/mean_length": 1143.875, "completions/mean_terminated_length": 893.6949462890625, "completions/min_length": 17.0, "completions/min_terminated_length": 17.0, "entropy": 3.6465570479631424, "epoch": 0.07936507936507936, "frac_reward_zero_std": 0.625, "grad_norm": 0.2353515625, "learning_rate": 9.21201814058957e-07, "loss": 0.0254, "num_tokens": 14082684.0, "reward": 0.1484375, "reward_std": 0.44257262349128723, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 140, "step_time": 287.1981354029849 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3880.0, "completions/mean_length": 1144.390625, "completions/mean_terminated_length": 722.732177734375, "completions/min_length": 136.0, "completions/min_terminated_length": 136.0, "entropy": 2.9444464705884457, "epoch": 0.07993197278911565, "frac_reward_zero_std": 0.75, "grad_norm": 0.181640625, "learning_rate": 9.206349206349205e-07, "loss": -0.0085, "num_tokens": 14166741.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 141, "step_time": 289.2533766971901 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3775.0, "completions/mean_length": 1429.53125, "completions/mean_terminated_length": 876.1132202148438, "completions/min_length": 33.0, "completions/min_terminated_length": 33.0, "entropy": 3.692947220057249, "epoch": 0.08049886621315193, "frac_reward_zero_std": 0.75, "grad_norm": 0.1689453125, "learning_rate": 9.200680272108843e-07, "loss": 0.0473, "num_tokens": 14269495.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 142, "step_time": 291.39080930687487 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4048.0, "completions/mean_length": 1398.734375, "completions/mean_terminated_length": 899.24072265625, "completions/min_length": 22.0, "completions/min_terminated_length": 22.0, "entropy": 3.200134973973036, "epoch": 0.0810657596371882, "frac_reward_zero_std": 0.75, "grad_norm": 0.1748046875, "learning_rate": 9.19501133786848e-07, "loss": 0.0324, "num_tokens": 14370518.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 143, "step_time": 292.93269517086446 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3933.0, "completions/mean_length": 1039.984375, "completions/mean_terminated_length": 781.0, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.9353726729750633, "epoch": 0.08163265306122448, "frac_reward_zero_std": 0.875, "grad_norm": 0.1943359375, "learning_rate": 9.189342403628118e-07, "loss": 0.0513, "num_tokens": 14448789.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 144, "step_time": 290.65329079981893 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3803.0, "completions/mean_length": 1379.59375, "completions/mean_terminated_length": 815.8113403320312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.1235422156751156, "epoch": 0.08219954648526077, "frac_reward_zero_std": 0.75, "grad_norm": 0.2470703125, "learning_rate": 9.183673469387755e-07, "loss": -0.0251, "num_tokens": 14550019.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 145, "step_time": 297.49194889795035 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3801.0, "completions/mean_length": 1014.796875, "completions/mean_terminated_length": 636.4035034179688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.78773495554924, "epoch": 0.08276643990929705, "frac_reward_zero_std": 0.625, "grad_norm": 0.28125, "learning_rate": 9.178004535147393e-07, "loss": -0.0003, "num_tokens": 14629438.0, "reward": 0.109375, "reward_std": 0.37234508991241455, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 146, "step_time": 307.780711280182 }, { "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": 0.25, "completions/max_length": 4096.0, "completions/max_terminated_length": 3994.0, "completions/mean_length": 1804.390625, "completions/mean_terminated_length": 1040.5208740234375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.545719593763351, "epoch": 0.08333333333333333, "frac_reward_zero_std": 0.75, "grad_norm": 0.1494140625, "learning_rate": 9.172335600907029e-07, "loss": 0.0256, "num_tokens": 14756431.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 147, "step_time": 291.5555063644424 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4036.0, "completions/mean_length": 1381.828125, "completions/mean_terminated_length": 755.4807739257812, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 2.881224237382412, "epoch": 0.08390022675736962, "frac_reward_zero_std": 0.75, "grad_norm": 0.2177734375, "learning_rate": 9.166666666666665e-07, "loss": 0.0029, "num_tokens": 14856028.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 148, "step_time": 289.2943117422983 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3943.0, "completions/mean_length": 1430.828125, "completions/mean_terminated_length": 937.2777709960938, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "entropy": 3.4363853074610233, "epoch": 0.0844671201814059, "frac_reward_zero_std": 0.75, "grad_norm": 0.1689453125, "learning_rate": 9.160997732426303e-07, "loss": 0.0258, "num_tokens": 14959313.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 149, "step_time": 291.73804178554565 }, { "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": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4031.0, "completions/mean_length": 1387.703125, "completions/mean_terminated_length": 697.3529663085938, "completions/min_length": 64.0, "completions/min_terminated_length": 64.0, "entropy": 3.1252135299146175, "epoch": 0.08503401360544217, "frac_reward_zero_std": 0.5, "grad_norm": 0.2578125, "learning_rate": 9.15532879818594e-07, "loss": -0.0041, "num_tokens": 15059574.0, "reward": 0.1953125, "reward_std": 0.5883605480194092, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 150, "step_time": 291.25094407517463 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3946.0, "completions/mean_length": 1650.984375, "completions/mean_terminated_length": 1143.5283203125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 4.018607679754496, "epoch": 0.08560090702947845, "frac_reward_zero_std": 0.875, "grad_norm": 0.1357421875, "learning_rate": 9.149659863945578e-07, "loss": 0.0087, "num_tokens": 15176845.0, "reward": 0.03125, "reward_std": 0.25, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 151, "step_time": 291.6673837006092 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4014.0, "completions/mean_length": 1243.4375, "completions/mean_terminated_length": 651.396240234375, "completions/min_length": 44.0, "completions/min_terminated_length": 44.0, "entropy": 2.930310718715191, "epoch": 0.08616780045351474, "frac_reward_zero_std": 0.875, "grad_norm": 0.1669921875, "learning_rate": 9.143990929705215e-07, "loss": 0.0313, "num_tokens": 15268545.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 152, "step_time": 292.3800488030538 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3808.0, "completions/mean_length": 1229.578125, "completions/mean_terminated_length": 760.5272216796875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.865586496889591, "epoch": 0.08673469387755102, "frac_reward_zero_std": 0.75, "grad_norm": 0.177734375, "learning_rate": 9.138321995464853e-07, "loss": -0.0067, "num_tokens": 15359238.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 153, "step_time": 290.51782300695777 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4047.0, "completions/mean_length": 1672.421875, "completions/mean_terminated_length": 1113.1346435546875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.8960447013378143, "epoch": 0.0873015873015873, "frac_reward_zero_std": 0.875, "grad_norm": 0.1181640625, "learning_rate": 9.132653061224489e-07, "loss": 0.0206, "num_tokens": 15477545.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 154, "step_time": 292.47963766381145 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3984.0, "completions/mean_length": 1202.875, "completions/mean_terminated_length": 789.5714721679688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.934021908789873, "epoch": 0.08786848072562359, "frac_reward_zero_std": 0.75, "grad_norm": 0.1806640625, "learning_rate": 9.126984126984127e-07, "loss": 0.0101, "num_tokens": 15570241.0, "reward": 0.1484375, "reward_std": 0.5092765688896179, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 155, "step_time": 309.80370240844786 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3906.0, "completions/mean_length": 1207.03125, "completions/mean_terminated_length": 794.3214721679688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.539234682917595, "epoch": 0.08843537414965986, "frac_reward_zero_std": 0.875, "grad_norm": 0.19921875, "learning_rate": 9.121315192743763e-07, "loss": -0.0083, "num_tokens": 15659643.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 156, "step_time": 295.2547142682597 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3429.0, "completions/mean_length": 1116.84375, "completions/mean_terminated_length": 864.3728637695312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.9506057165563107, "epoch": 0.08900226757369614, "frac_reward_zero_std": 0.5, "grad_norm": 0.3515625, "learning_rate": 9.115646258503401e-07, "loss": 0.005, "num_tokens": 15742537.0, "reward": 0.171875, "reward_std": 0.5434950590133667, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 157, "step_time": 289.8194957161322 }, { "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": 0.03125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3172.0, "completions/mean_length": 717.765625, "completions/mean_terminated_length": 608.790283203125, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.404065862298012, "epoch": 0.08956916099773243, "frac_reward_zero_std": 0.75, "grad_norm": 0.376953125, "learning_rate": 9.109977324263038e-07, "loss": 0.0665, "num_tokens": 15798634.0, "reward": 0.15625, "reward_std": 0.5261042714118958, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 158, "step_time": 286.81534686125815 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3212.0, "completions/mean_length": 1162.484375, "completions/mean_terminated_length": 802.2280883789062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.865950843319297, "epoch": 0.09013605442176871, "frac_reward_zero_std": 0.75, "grad_norm": 0.23828125, "learning_rate": 9.104308390022676e-07, "loss": 0.0076, "num_tokens": 15885585.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 159, "step_time": 294.96963218413293 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4079.0, "completions/mean_length": 1392.46875, "completions/mean_terminated_length": 831.3585205078125, "completions/min_length": 144.0, "completions/min_terminated_length": 144.0, "entropy": 3.362879954278469, "epoch": 0.09070294784580499, "frac_reward_zero_std": 0.875, "grad_norm": 0.1171875, "learning_rate": 9.098639455782312e-07, "loss": 0.0177, "num_tokens": 15985199.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 160, "step_time": 287.4931676927954 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3987.0, "completions/mean_length": 1498.546875, "completions/mean_terminated_length": 1017.5370483398438, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.24854800850153, "epoch": 0.09126984126984126, "frac_reward_zero_std": 0.75, "grad_norm": 0.20703125, "learning_rate": 9.09297052154195e-07, "loss": 0.0382, "num_tokens": 16093890.0, "reward": 0.125, "reward_std": 0.4364357888698578, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 161, "step_time": 294.40642016381025 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3951.0, "completions/mean_length": 1150.96875, "completions/mean_terminated_length": 846.3103637695312, "completions/min_length": 69.0, "completions/min_terminated_length": 69.0, "entropy": 3.4237146079540253, "epoch": 0.09183673469387756, "frac_reward_zero_std": 0.625, "grad_norm": 0.25390625, "learning_rate": 9.087301587301587e-07, "loss": 0.0352, "num_tokens": 16178136.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 162, "step_time": 288.3448597891256 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3924.0, "completions/mean_length": 1490.609375, "completions/mean_terminated_length": 1008.129638671875, "completions/min_length": 235.0, "completions/min_terminated_length": 235.0, "entropy": 3.4717009142041206, "epoch": 0.09240362811791383, "frac_reward_zero_std": 0.75, "grad_norm": 0.1630859375, "learning_rate": 9.081632653061225e-07, "loss": 0.0173, "num_tokens": 16285743.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 163, "step_time": 291.6983943702653 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3994.0, "completions/mean_length": 1456.28125, "completions/mean_terminated_length": 967.4444580078125, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "entropy": 3.486091446131468, "epoch": 0.09297052154195011, "frac_reward_zero_std": 0.75, "grad_norm": 0.248046875, "learning_rate": 9.075963718820861e-07, "loss": -0.0448, "num_tokens": 16392361.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 164, "step_time": 294.49498210661113 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3656.0, "completions/mean_length": 1174.640625, "completions/mean_terminated_length": 757.3035888671875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.4663752168416977, "epoch": 0.0935374149659864, "frac_reward_zero_std": 0.75, "grad_norm": 0.189453125, "learning_rate": 9.070294784580499e-07, "loss": -0.0097, "num_tokens": 16477842.0, "reward": 0.078125, "reward_std": 0.2847827076911926, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 165, "step_time": 289.076499841176 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3888.0, "completions/mean_length": 1292.578125, "completions/mean_terminated_length": 892.08935546875, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 3.132295398041606, "epoch": 0.09410430839002268, "frac_reward_zero_std": 0.875, "grad_norm": 0.16796875, "learning_rate": 9.064625850340135e-07, "loss": 0.0079, "num_tokens": 16572823.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 166, "step_time": 299.84396067727357 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3867.0, "completions/mean_length": 1318.265625, "completions/mean_terminated_length": 803.870361328125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.2404509000480175, "epoch": 0.09467120181405896, "frac_reward_zero_std": 0.875, "grad_norm": 0.115234375, "learning_rate": 9.058956916099772e-07, "loss": -0.0171, "num_tokens": 16668464.0, "reward": 0.171875, "reward_std": 0.5579067468643188, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 167, "step_time": 294.01645047590137 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 2921.0, "completions/mean_length": 1203.375, "completions/mean_terminated_length": 667.7037353515625, "completions/min_length": 39.0, "completions/min_terminated_length": 39.0, "entropy": 2.8669565990567207, "epoch": 0.09523809523809523, "frac_reward_zero_std": 0.75, "grad_norm": 0.2255859375, "learning_rate": 9.05328798185941e-07, "loss": 0.0009, "num_tokens": 16758040.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 168, "step_time": 294.837905542925 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3700.0, "completions/mean_length": 1264.515625, "completions/mean_terminated_length": 1024.559326171875, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.5907875560224056, "epoch": 0.09580498866213152, "frac_reward_zero_std": 0.875, "grad_norm": 0.1435546875, "learning_rate": 9.047619047619047e-07, "loss": 0.007, "num_tokens": 16851593.0, "reward": 0.03125, "reward_std": 0.25, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 169, "step_time": 292.3823089450598 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3599.0, "completions/mean_length": 1052.8125, "completions/mean_terminated_length": 618.0714721679688, "completions/min_length": 19.0, "completions/min_terminated_length": 19.0, "entropy": 2.5225942097604275, "epoch": 0.0963718820861678, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 9.041950113378685e-07, "loss": 0.0, "num_tokens": 16933941.0, "reward": 0.0625, "reward_std": 0.1666666716337204, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 170, "step_time": 308.9409235548228 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4080.0, "completions/mean_length": 1429.15625, "completions/mean_terminated_length": 935.2963256835938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.53740531206131, "epoch": 0.09693877551020408, "frac_reward_zero_std": 0.625, "grad_norm": 0.22265625, "learning_rate": 9.036281179138322e-07, "loss": 0.0408, "num_tokens": 17035623.0, "reward": 0.109375, "reward_std": 0.4315783977508545, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 171, "step_time": 287.8768406538293 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3887.0, "completions/mean_length": 1371.15625, "completions/mean_terminated_length": 1036.5263671875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 5.13199670612812, "epoch": 0.09750566893424037, "frac_reward_zero_std": 0.75, "grad_norm": 0.1904296875, "learning_rate": 9.030612244897958e-07, "loss": 0.0186, "num_tokens": 17134569.0, "reward": 0.140625, "reward_std": 0.4917473793029785, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 172, "step_time": 292.4301679106429 }, { "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": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3977.0, "completions/mean_length": 1616.921875, "completions/mean_terminated_length": 922.7799682617188, "completions/min_length": 92.0, "completions/min_terminated_length": 92.0, "entropy": 3.5396290235221386, "epoch": 0.09807256235827665, "frac_reward_zero_std": 0.625, "grad_norm": 0.166015625, "learning_rate": 9.024943310657595e-07, "loss": 0.0189, "num_tokens": 17249228.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 173, "step_time": 289.1016967650503 }, { "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": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3787.0, "completions/mean_length": 1590.9375, "completions/mean_terminated_length": 952.3922119140625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.543005548417568, "epoch": 0.09863945578231292, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 9.019274376417233e-07, "loss": 0.0, "num_tokens": 17362840.0, "reward": 0.0, "reward_std": 0.0, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 174, "step_time": 292.75521798059344 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3958.0, "completions/mean_length": 1500.03125, "completions/mean_terminated_length": 961.2453002929688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.7964779548346996, "epoch": 0.0992063492063492, "frac_reward_zero_std": 0.75, "grad_norm": 0.1767578125, "learning_rate": 9.01360544217687e-07, "loss": 0.0297, "num_tokens": 17472946.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 175, "step_time": 311.0415435722098 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 2953.0, "completions/mean_length": 1142.703125, "completions/mean_terminated_length": 720.8035888671875, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 2.593105688691139, "epoch": 0.09977324263038549, "frac_reward_zero_std": 0.75, "grad_norm": 0.205078125, "learning_rate": 9.007936507936508e-07, "loss": -0.0259, "num_tokens": 17557383.0, "reward": 0.140625, "reward_std": 0.4406772553920746, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 176, "step_time": 290.1869981242344 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3650.0, "completions/mean_length": 1075.359375, "completions/mean_terminated_length": 762.8793334960938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 1.4477510415017605, "epoch": 0.10034013605442177, "frac_reward_zero_std": 0.75, "grad_norm": 0.1962890625, "learning_rate": 9.002267573696145e-07, "loss": 0.0103, "num_tokens": 17637182.0, "reward": 0.1015625, "reward_std": 0.3692515790462494, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 177, "step_time": 290.2055966453627 }, { "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": 4096.0, "completions/max_terminated_length": 3889.0, "completions/mean_length": 1227.375, "completions/mean_terminated_length": 1036.1334228515625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.749495092779398, "epoch": 0.10090702947845805, "frac_reward_zero_std": 0.75, "grad_norm": 0.2421875, "learning_rate": 8.996598639455783e-07, "loss": 0.0163, "num_tokens": 17726174.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 178, "step_time": 291.8939747372642 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3426.0, "completions/mean_length": 1324.1875, "completions/mean_terminated_length": 928.21435546875, "completions/min_length": 64.0, "completions/min_terminated_length": 64.0, "entropy": 3.1194748021662235, "epoch": 0.10147392290249434, "frac_reward_zero_std": 0.75, "grad_norm": 0.232421875, "learning_rate": 8.990929705215419e-07, "loss": -0.0212, "num_tokens": 17823706.0, "reward": 0.1015625, "reward_std": 0.40999117493629456, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 179, "step_time": 300.7299531698227 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3539.0, "completions/mean_length": 1369.484375, "completions/mean_terminated_length": 979.982177734375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.87375595793128, "epoch": 0.10204081632653061, "frac_reward_zero_std": 0.875, "grad_norm": 0.1064453125, "learning_rate": 8.985260770975056e-07, "loss": 0.0002, "num_tokens": 17925425.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 180, "step_time": 309.0694700097665 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4075.0, "completions/mean_length": 1371.796875, "completions/mean_terminated_length": 867.3148193359375, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 4.180113475769758, "epoch": 0.10260770975056689, "frac_reward_zero_std": 0.875, "grad_norm": 0.11376953125, "learning_rate": 8.979591836734693e-07, "loss": -0.0025, "num_tokens": 18025684.0, "reward": 0.0546875, "reward_std": 0.15728822350502014, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 181, "step_time": 296.59269960038364 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3964.0, "completions/mean_length": 1226.59375, "completions/mean_terminated_length": 874.2105102539062, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.224886655807495, "epoch": 0.10317460317460317, "frac_reward_zero_std": 0.75, "grad_norm": 0.2099609375, "learning_rate": 8.973922902494331e-07, "loss": 0.0103, "num_tokens": 18115266.0, "reward": 0.15625, "reward_std": 0.49501484632492065, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 182, "step_time": 289.9309525163844 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 2623.0, "completions/mean_length": 1054.765625, "completions/mean_terminated_length": 681.2807006835938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.514544364064932, "epoch": 0.10374149659863946, "frac_reward_zero_std": 0.875, "grad_norm": 0.2421875, "learning_rate": 8.968253968253968e-07, "loss": 0.0059, "num_tokens": 18193963.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 183, "step_time": 290.2435266915709 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3834.0, "completions/mean_length": 1142.796875, "completions/mean_terminated_length": 892.5254516601562, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "entropy": 3.2445945367217064, "epoch": 0.10430839002267574, "frac_reward_zero_std": 0.625, "grad_norm": 0.3125, "learning_rate": 8.962585034013606e-07, "loss": 0.0091, "num_tokens": 18278142.0, "reward": 0.15625, "reward_std": 0.5261042714118958, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 184, "step_time": 288.83665825333446 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3869.0, "completions/mean_length": 1482.265625, "completions/mean_terminated_length": 1054.5635986328125, "completions/min_length": 110.0, "completions/min_terminated_length": 110.0, "entropy": 2.9784867502748966, "epoch": 0.10487528344671201, "frac_reward_zero_std": 0.5, "grad_norm": 0.318359375, "learning_rate": 8.956916099773242e-07, "loss": 0.0083, "num_tokens": 18383759.0, "reward": 0.1328125, "reward_std": 0.489915668964386, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 185, "step_time": 291.71841166727245 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4081.0, "completions/mean_length": 1335.828125, "completions/mean_terminated_length": 941.5178833007812, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "entropy": 3.1029345877468586, "epoch": 0.1054421768707483, "frac_reward_zero_std": 0.875, "grad_norm": 0.12890625, "learning_rate": 8.95124716553288e-07, "loss": 0.0155, "num_tokens": 18481956.0, "reward": 0.03125, "reward_std": 0.25, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 186, "step_time": 296.8026285711676 }, { "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": 0.28125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3717.0, "completions/mean_length": 1743.265625, "completions/mean_terminated_length": 822.6304321289062, "completions/min_length": 145.0, "completions/min_terminated_length": 145.0, "entropy": 2.540066074579954, "epoch": 0.10600907029478458, "frac_reward_zero_std": 0.625, "grad_norm": 0.1474609375, "learning_rate": 8.945578231292517e-07, "loss": -0.0017, "num_tokens": 18604453.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 187, "step_time": 288.1224105041474 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 2678.0, "completions/mean_length": 1222.453125, "completions/mean_terminated_length": 690.3148193359375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 1.7623973712325096, "epoch": 0.10657596371882086, "frac_reward_zero_std": 0.5, "grad_norm": 0.35546875, "learning_rate": 8.939909297052153e-07, "loss": -0.0119, "num_tokens": 18694050.0, "reward": 0.1328125, "reward_std": 0.489915668964386, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 188, "step_time": 289.81556538585573 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3911.0, "completions/mean_length": 1203.671875, "completions/mean_terminated_length": 848.4736938476562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.732121203094721, "epoch": 0.10714285714285714, "frac_reward_zero_std": 0.75, "grad_norm": 0.2236328125, "learning_rate": 8.934240362811791e-07, "loss": 0.0008, "num_tokens": 18784061.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 189, "step_time": 294.01357916835696 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3785.0, "completions/mean_length": 1114.328125, "completions/mean_terminated_length": 967.6884765625, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 4.224556617438793, "epoch": 0.10770975056689343, "frac_reward_zero_std": 0.75, "grad_norm": 0.2294921875, "learning_rate": 8.928571428571428e-07, "loss": 0.0132, "num_tokens": 18866546.0, "reward": 0.125, "reward_std": 0.4364357888698578, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 190, "step_time": 292.27853817585856 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3553.0, "completions/mean_length": 1233.328125, "completions/mean_terminated_length": 824.3750610351562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.003290619701147, "epoch": 0.1082766439909297, "frac_reward_zero_std": 0.5, "grad_norm": 0.3125, "learning_rate": 8.922902494331065e-07, "loss": 0.0651, "num_tokens": 18959071.0, "reward": 0.1484375, "reward_std": 0.49344661831855774, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 191, "step_time": 296.30524972453713 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 2366.0, "completions/mean_length": 898.953125, "completions/mean_terminated_length": 628.0169677734375, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "entropy": 2.3804745860397816, "epoch": 0.10884353741496598, "frac_reward_zero_std": 0.875, "grad_norm": 0.20703125, "learning_rate": 8.917233560090702e-07, "loss": 0.0131, "num_tokens": 19027940.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 192, "step_time": 290.5428119143471 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3684.0, "completions/mean_length": 930.0625, "completions/mean_terminated_length": 661.7626953125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.6110351644456387, "epoch": 0.10941043083900227, "frac_reward_zero_std": 0.875, "grad_norm": 0.255859375, "learning_rate": 8.91156462585034e-07, "loss": -0.022, "num_tokens": 19098552.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 193, "step_time": 292.44580878596753 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3931.0, "completions/mean_length": 1209.875, "completions/mean_terminated_length": 965.2881469726562, "completions/min_length": 115.0, "completions/min_terminated_length": 115.0, "entropy": 2.891344264149666, "epoch": 0.10997732426303855, "frac_reward_zero_std": 0.875, "grad_norm": 0.1845703125, "learning_rate": 8.905895691609977e-07, "loss": -0.0063, "num_tokens": 19186168.0, "reward": 0.046875, "reward_std": 0.14689241349697113, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 194, "step_time": 288.53979091718793 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4035.0, "completions/mean_length": 1533.671875, "completions/mean_terminated_length": 1219.0, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 5.731172285974026, "epoch": 0.11054421768707483, "frac_reward_zero_std": 0.625, "grad_norm": 0.291015625, "learning_rate": 8.900226757369615e-07, "loss": -0.0531, "num_tokens": 19294899.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 195, "step_time": 290.09615557082 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3942.0, "completions/mean_length": 1429.15625, "completions/mean_terminated_length": 935.2963256835938, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "entropy": 4.4110739547759295, "epoch": 0.1111111111111111, "frac_reward_zero_std": 0.875, "grad_norm": 0.1357421875, "learning_rate": 8.894557823129251e-07, "loss": 0.0339, "num_tokens": 19401493.0, "reward": 0.046875, "reward_std": 0.14689241349697113, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 196, "step_time": 321.8842496164143 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3876.0, "completions/mean_length": 1081.484375, "completions/mean_terminated_length": 769.637939453125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.056702673435211, "epoch": 0.1116780045351474, "frac_reward_zero_std": 0.75, "grad_norm": 0.2294921875, "learning_rate": 8.888888888888888e-07, "loss": -0.0012, "num_tokens": 19484276.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 197, "step_time": 297.40092708636075 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3986.0, "completions/mean_length": 1091.421875, "completions/mean_terminated_length": 780.6034545898438, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.2152112014591694, "epoch": 0.11224489795918367, "frac_reward_zero_std": 0.625, "grad_norm": 0.326171875, "learning_rate": 8.883219954648525e-07, "loss": 0.0865, "num_tokens": 19566455.0, "reward": 0.1171875, "reward_std": 0.37524792551994324, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 198, "step_time": 298.85880855005234 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4029.0, "completions/mean_length": 1032.765625, "completions/mean_terminated_length": 882.1146850585938, "completions/min_length": 84.0, "completions/min_terminated_length": 84.0, "entropy": 3.6219312641769648, "epoch": 0.11281179138321995, "frac_reward_zero_std": 0.875, "grad_norm": 0.220703125, "learning_rate": 8.877551020408163e-07, "loss": 0.0288, "num_tokens": 19643024.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 199, "step_time": 290.5095385769382 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4055.0, "completions/mean_length": 1387.609375, "completions/mean_terminated_length": 886.0555419921875, "completions/min_length": 31.0, "completions/min_terminated_length": 31.0, "entropy": 4.273848362267017, "epoch": 0.11337868480725624, "frac_reward_zero_std": 0.875, "grad_norm": 0.1396484375, "learning_rate": 8.8718820861678e-07, "loss": 0.0349, "num_tokens": 19743087.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 200, "step_time": 289.47388445865363 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3824.0, "completions/mean_length": 1406.453125, "completions/mean_terminated_length": 1022.232177734375, "completions/min_length": 24.0, "completions/min_terminated_length": 24.0, "entropy": 4.070297235623002, "epoch": 0.11394557823129252, "frac_reward_zero_std": 0.75, "grad_norm": 0.1845703125, "learning_rate": 8.866213151927438e-07, "loss": 0.0337, "num_tokens": 19844340.0, "reward": 0.1875, "reward_std": 0.5875696539878845, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 201, "step_time": 291.9551081676036 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3642.0, "completions/mean_length": 1340.34375, "completions/mean_terminated_length": 889.4181518554688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.479163069278002, "epoch": 0.1145124716553288, "frac_reward_zero_std": 0.625, "grad_norm": 0.28515625, "learning_rate": 8.860544217687075e-07, "loss": 0.0271, "num_tokens": 19943482.0, "reward": 0.1796875, "reward_std": 0.5445777773857117, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 202, "step_time": 307.04083415772766 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3761.0, "completions/mean_length": 1303.234375, "completions/mean_terminated_length": 786.0555419921875, "completions/min_length": 77.0, "completions/min_terminated_length": 77.0, "entropy": 2.687175914645195, "epoch": 0.11507936507936507, "frac_reward_zero_std": 0.875, "grad_norm": 0.15625, "learning_rate": 8.854875283446712e-07, "loss": 0.0109, "num_tokens": 20037729.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 203, "step_time": 290.80970280617476 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3968.0, "completions/mean_length": 1383.078125, "completions/mean_terminated_length": 995.5178833007812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 4.320884607732296, "epoch": 0.11564625850340136, "frac_reward_zero_std": 0.75, "grad_norm": 0.203125, "learning_rate": 8.849206349206348e-07, "loss": 0.0488, "num_tokens": 20137494.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 204, "step_time": 291.97664872650057 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3204.0, "completions/mean_length": 1105.53125, "completions/mean_terminated_length": 678.3214721679688, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.0910265929996967, "epoch": 0.11621315192743764, "frac_reward_zero_std": 0.375, "grad_norm": 0.373046875, "learning_rate": 8.843537414965985e-07, "loss": 0.0423, "num_tokens": 20218472.0, "reward": 0.15625, "reward_std": 0.49501484632492065, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 205, "step_time": 287.7204177901149 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3842.0, "completions/mean_length": 1305.875, "completions/mean_terminated_length": 1017.2413940429688, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "entropy": 3.4122538827359676, "epoch": 0.11678004535147392, "frac_reward_zero_std": 0.875, "grad_norm": 0.1376953125, "learning_rate": 8.837868480725623e-07, "loss": 0.0119, "num_tokens": 20315120.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 206, "step_time": 296.36069345939904 }, { "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": 0.03125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3730.0, "completions/mean_length": 890.453125, "completions/mean_terminated_length": 787.04833984375, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "entropy": 3.4672073535621166, "epoch": 0.11734693877551021, "frac_reward_zero_std": 0.5, "grad_norm": 0.322265625, "learning_rate": 8.83219954648526e-07, "loss": 0.0688, "num_tokens": 20384789.0, "reward": 0.28125, "reward_std": 0.6538955569267273, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 207, "step_time": 304.66316656395793 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3863.0, "completions/mean_length": 1325.53125, "completions/mean_terminated_length": 812.4815063476562, "completions/min_length": 160.0, "completions/min_terminated_length": 160.0, "entropy": 2.7345456182956696, "epoch": 0.11791383219954649, "frac_reward_zero_std": 0.875, "grad_norm": 0.1552734375, "learning_rate": 8.826530612244898e-07, "loss": 0.0174, "num_tokens": 20480623.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 208, "step_time": 290.17180477362126 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3560.0, "completions/mean_length": 1259.1875, "completions/mean_terminated_length": 794.9818115234375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.216690521687269, "epoch": 0.11848072562358276, "frac_reward_zero_std": 0.875, "grad_norm": 0.169921875, "learning_rate": 8.820861678004535e-07, "loss": 0.0237, "num_tokens": 20571763.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 209, "step_time": 287.59152660612017 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4094.0, "completions/mean_length": 1180.515625, "completions/mean_terminated_length": 878.913818359375, "completions/min_length": 17.0, "completions/min_terminated_length": 17.0, "entropy": 3.4672042913734913, "epoch": 0.11904761904761904, "frac_reward_zero_std": 0.75, "grad_norm": 0.2353515625, "learning_rate": 8.815192743764172e-07, "loss": 0.0128, "num_tokens": 20658684.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 210, "step_time": 288.88174922112375 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3385.0, "completions/mean_length": 1296.046875, "completions/mean_terminated_length": 896.0535888671875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.265394568443298, "epoch": 0.11961451247165533, "frac_reward_zero_std": 0.75, "grad_norm": 0.26171875, "learning_rate": 8.809523809523809e-07, "loss": -0.0064, "num_tokens": 20752759.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 211, "step_time": 292.75100475084037 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3691.0, "completions/mean_length": 1512.234375, "completions/mean_terminated_length": 1033.75927734375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.9343014881014824, "epoch": 0.12018140589569161, "frac_reward_zero_std": 0.875, "grad_norm": 0.1572265625, "learning_rate": 8.803854875283446e-07, "loss": 0.0243, "num_tokens": 20860598.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 212, "step_time": 291.02801759541035 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3392.0, "completions/mean_length": 1100.4375, "completions/mean_terminated_length": 846.5762939453125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.2256470173597336, "epoch": 0.12074829931972789, "frac_reward_zero_std": 0.875, "grad_norm": 0.11572265625, "learning_rate": 8.798185941043083e-07, "loss": -0.0028, "num_tokens": 20943858.0, "reward": 0.0546875, "reward_std": 0.15728822350502014, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 213, "step_time": 294.07494006305933 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4044.0, "completions/mean_length": 1390.53125, "completions/mean_terminated_length": 1058.28076171875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.787644349038601, "epoch": 0.12131519274376418, "frac_reward_zero_std": 0.75, "grad_norm": 0.28125, "learning_rate": 8.792517006802721e-07, "loss": 0.0849, "num_tokens": 21044732.0, "reward": 0.0859375, "reward_std": 0.28942590951919556, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 214, "step_time": 296.0653443885967 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3992.0, "completions/mean_length": 1385.5625, "completions/mean_terminated_length": 883.629638671875, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "entropy": 3.013167019933462, "epoch": 0.12188208616780045, "frac_reward_zero_std": 0.75, "grad_norm": 0.1552734375, "learning_rate": 8.786848072562358e-07, "loss": -0.0229, "num_tokens": 21145048.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 215, "step_time": 294.7712307209149 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3962.0, "completions/mean_length": 1490.28125, "completions/mean_terminated_length": 888.9615478515625, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "entropy": 3.3624575063586235, "epoch": 0.12244897959183673, "frac_reward_zero_std": 0.75, "grad_norm": 0.193359375, "learning_rate": 8.781179138321995e-07, "loss": -0.0368, "num_tokens": 21256330.0, "reward": 0.1171875, "reward_std": 0.43408530950546265, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 216, "step_time": 309.2507579885423 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3956.0, "completions/mean_length": 1153.75, "completions/mean_terminated_length": 849.3793334960938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.0180613733828068, "epoch": 0.12301587301587301, "frac_reward_zero_std": 0.75, "grad_norm": 0.30078125, "learning_rate": 8.775510204081632e-07, "loss": -0.003, "num_tokens": 21341482.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 217, "step_time": 296.18976107239723 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4088.0, "completions/mean_length": 1482.5625, "completions/mean_terminated_length": 1109.21435546875, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 4.036535518243909, "epoch": 0.1235827664399093, "frac_reward_zero_std": 0.875, "grad_norm": 0.1630859375, "learning_rate": 8.76984126984127e-07, "loss": 0.0172, "num_tokens": 21448398.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 218, "step_time": 294.8545334087685 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3274.0, "completions/mean_length": 1120.59375, "completions/mean_terminated_length": 633.7090454101562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.101956117898226, "epoch": 0.12414965986394558, "frac_reward_zero_std": 0.75, "grad_norm": 0.1953125, "learning_rate": 8.764172335600907e-07, "loss": -0.0004, "num_tokens": 21531412.0, "reward": 0.1015625, "reward_std": 0.3692515790462494, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 219, "step_time": 287.65193302184343 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3703.0, "completions/mean_length": 1089.578125, "completions/mean_terminated_length": 778.5689697265625, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "entropy": 2.9111460521817207, "epoch": 0.12471655328798185, "frac_reward_zero_std": 0.875, "grad_norm": 0.212890625, "learning_rate": 8.758503401360544e-07, "loss": 0.0203, "num_tokens": 21613137.0, "reward": 0.046875, "reward_std": 0.14689241349697113, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 220, "step_time": 293.88446793332696 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3863.0, "completions/mean_length": 1216.734375, "completions/mean_terminated_length": 918.8793334960938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.4708727560937405, "epoch": 0.12528344671201813, "frac_reward_zero_std": 0.75, "grad_norm": 0.33984375, "learning_rate": 8.752834467120181e-07, "loss": 0.0417, "num_tokens": 21702936.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 221, "step_time": 289.8535577338189 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4002.0, "completions/mean_length": 1150.015625, "completions/mean_terminated_length": 845.2586059570312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.823819410055876, "epoch": 0.12585034013605442, "frac_reward_zero_std": 0.625, "grad_norm": 0.3359375, "learning_rate": 8.747165532879818e-07, "loss": 0.0275, "num_tokens": 21790105.0, "reward": 0.1484375, "reward_std": 0.44257262349128723, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 222, "step_time": 307.9171469854191 }, { "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": 4096.0, "completions/max_terminated_length": 3438.0, "completions/mean_length": 904.375, "completions/mean_terminated_length": 691.6000366210938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.9799317941069603, "epoch": 0.1264172335600907, "frac_reward_zero_std": 0.625, "grad_norm": 0.36328125, "learning_rate": 8.741496598639455e-07, "loss": 0.0336, "num_tokens": 21858849.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 223, "step_time": 288.9901874726638 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3831.0, "completions/mean_length": 1087.34375, "completions/mean_terminated_length": 717.8596801757812, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "entropy": 2.7648992855101824, "epoch": 0.12698412698412698, "frac_reward_zero_std": 0.625, "grad_norm": 0.2490234375, "learning_rate": 8.735827664399093e-07, "loss": 0.0314, "num_tokens": 21939375.0, "reward": 0.1328125, "reward_std": 0.489915668964386, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 224, "step_time": 288.9207858527079 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3380.0, "completions/mean_length": 946.53125, "completions/mean_terminated_length": 679.6271362304688, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "entropy": 2.9799016043543816, "epoch": 0.12755102040816327, "frac_reward_zero_std": 0.625, "grad_norm": 0.345703125, "learning_rate": 8.73015873015873e-07, "loss": 0.0557, "num_tokens": 22017185.0, "reward": 0.2109375, "reward_std": 0.6029338836669922, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 225, "step_time": 333.07871426828206 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3114.0, "completions/mean_length": 997.484375, "completions/mean_terminated_length": 676.9483032226562, "completions/min_length": 70.0, "completions/min_terminated_length": 70.0, "entropy": 2.2997788041830063, "epoch": 0.12811791383219956, "frac_reward_zero_std": 0.625, "grad_norm": 0.28515625, "learning_rate": 8.724489795918368e-07, "loss": 0.0429, "num_tokens": 22091840.0, "reward": 0.140625, "reward_std": 0.4917473793029785, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 226, "step_time": 290.95938858762383 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3887.0, "completions/mean_length": 1171.5625, "completions/mean_terminated_length": 923.7288208007812, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.4151121973991394, "epoch": 0.12868480725623582, "frac_reward_zero_std": 0.875, "grad_norm": 0.197265625, "learning_rate": 8.718820861678005e-07, "loss": 0.0418, "num_tokens": 22178036.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 227, "step_time": 292.3845506478101 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4006.0, "completions/mean_length": 1215.59375, "completions/mean_terminated_length": 744.2545166015625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.0569019466638565, "epoch": 0.1292517006802721, "frac_reward_zero_std": 0.75, "grad_norm": 0.1884765625, "learning_rate": 8.71315192743764e-07, "loss": 0.0042, "num_tokens": 22266506.0, "reward": 0.171875, "reward_std": 0.5434950590133667, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 228, "step_time": 288.30535659752786 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3954.0, "completions/mean_length": 1202.421875, "completions/mean_terminated_length": 957.203369140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.267720803618431, "epoch": 0.12981859410430838, "frac_reward_zero_std": 0.75, "grad_norm": 0.234375, "learning_rate": 8.707482993197278e-07, "loss": -0.0422, "num_tokens": 22354677.0, "reward": 0.0625, "reward_std": 0.35073620080947876, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 229, "step_time": 291.8245045123622 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3983.0, "completions/mean_length": 1506.796875, "completions/mean_terminated_length": 1027.3148193359375, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "entropy": 4.22579360473901, "epoch": 0.13038548752834467, "frac_reward_zero_std": 0.875, "grad_norm": 0.126953125, "learning_rate": 8.701814058956915e-07, "loss": -0.0085, "num_tokens": 22462576.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 230, "step_time": 291.576286659576 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4041.0, "completions/mean_length": 1392.15625, "completions/mean_terminated_length": 949.7090454101562, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.5622212290763855, "epoch": 0.13095238095238096, "frac_reward_zero_std": 0.75, "grad_norm": 0.224609375, "learning_rate": 8.696145124716553e-07, "loss": -0.0263, "num_tokens": 22561978.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 231, "step_time": 290.1018397184089 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3572.0, "completions/mean_length": 1307.28125, "completions/mean_terminated_length": 850.9454345703125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.2057136446237564, "epoch": 0.13151927437641722, "frac_reward_zero_std": 0.375, "grad_norm": 0.28515625, "learning_rate": 8.69047619047619e-07, "loss": 0.0917, "num_tokens": 22657732.0, "reward": 0.2109375, "reward_std": 0.5896238088607788, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 232, "step_time": 293.98524337727576 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3860.0, "completions/mean_length": 1025.15625, "completions/mean_terminated_length": 874.131103515625, "completions/min_length": 67.0, "completions/min_terminated_length": 67.0, "entropy": 4.575048711150885, "epoch": 0.1320861678004535, "frac_reward_zero_std": 0.75, "grad_norm": 0.2109375, "learning_rate": 8.684807256235828e-07, "loss": 0.0255, "num_tokens": 22734878.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 233, "step_time": 290.87545877788216 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3896.0, "completions/mean_length": 1130.6875, "completions/mean_terminated_length": 879.3898315429688, "completions/min_length": 152.0, "completions/min_terminated_length": 152.0, "entropy": 3.9855271950364113, "epoch": 0.1326530612244898, "frac_reward_zero_std": 0.75, "grad_norm": 0.259765625, "learning_rate": 8.679138321995464e-07, "loss": -0.0024, "num_tokens": 22819754.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 234, "step_time": 299.91231037303805 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3671.0, "completions/mean_length": 1123.328125, "completions/mean_terminated_length": 871.4067993164062, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "entropy": 4.110420741140842, "epoch": 0.13321995464852607, "frac_reward_zero_std": 0.875, "grad_norm": 0.1669921875, "learning_rate": 8.673469387755102e-07, "loss": 0.0482, "num_tokens": 22907759.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 235, "step_time": 311.0638914471492 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4078.0, "completions/mean_length": 1541.078125, "completions/mean_terminated_length": 1067.9444580078125, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "entropy": 3.5040894635021687, "epoch": 0.13378684807256236, "frac_reward_zero_std": 0.75, "grad_norm": 0.203125, "learning_rate": 8.667800453514738e-07, "loss": 0.0899, "num_tokens": 23021684.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 236, "step_time": 304.27577920630574 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3680.0, "completions/mean_length": 1308.171875, "completions/mean_terminated_length": 791.9074096679688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.26109742000699, "epoch": 0.13435374149659865, "frac_reward_zero_std": 0.875, "grad_norm": 0.1875, "learning_rate": 8.662131519274376e-07, "loss": -0.0342, "num_tokens": 23117255.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 237, "step_time": 296.1404558233917 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3953.0, "completions/mean_length": 1286.234375, "completions/mean_terminated_length": 884.83935546875, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "entropy": 3.5244238078594208, "epoch": 0.1349206349206349, "frac_reward_zero_std": 0.75, "grad_norm": 0.267578125, "learning_rate": 8.656462585034013e-07, "loss": -0.0215, "num_tokens": 23210974.0, "reward": 0.1015625, "reward_std": 0.42891231179237366, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 238, "step_time": 290.48089622799307 }, { "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": 4096.0, "completions/max_terminated_length": 3929.0, "completions/mean_length": 1086.921875, "completions/mean_terminated_length": 886.3167114257812, "completions/min_length": 187.0, "completions/min_terminated_length": 187.0, "entropy": 2.5971093960106373, "epoch": 0.1354875283446712, "frac_reward_zero_std": 0.875, "grad_norm": 0.220703125, "learning_rate": 8.650793650793651e-07, "loss": 0.0309, "num_tokens": 23291625.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 239, "step_time": 291.6085400208831 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3466.0, "completions/mean_length": 1168.3125, "completions/mean_terminated_length": 865.4482421875, "completions/min_length": 26.0, "completions/min_terminated_length": 26.0, "entropy": 3.861044980585575, "epoch": 0.1360544217687075, "frac_reward_zero_std": 0.75, "grad_norm": 0.2275390625, "learning_rate": 8.645124716553288e-07, "loss": 0.053, "num_tokens": 23377477.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 240, "step_time": 290.8934416761622 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3898.0, "completions/mean_length": 1449.78125, "completions/mean_terminated_length": 839.1154174804688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.9521565549075603, "epoch": 0.13662131519274376, "frac_reward_zero_std": 0.625, "grad_norm": 0.205078125, "learning_rate": 8.639455782312925e-07, "loss": 0.0162, "num_tokens": 23481647.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 241, "step_time": 292.27017205022275 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 2670.0, "completions/mean_length": 1245.359375, "completions/mean_terminated_length": 717.4629516601562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.227583382278681, "epoch": 0.13718820861678005, "frac_reward_zero_std": 0.75, "grad_norm": 0.1630859375, "learning_rate": 8.633786848072562e-07, "loss": 0.0255, "num_tokens": 23572566.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 242, "step_time": 292.54244574252516 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 2453.0, "completions/mean_length": 832.84375, "completions/mean_terminated_length": 495.2758483886719, "completions/min_length": 114.0, "completions/min_terminated_length": 114.0, "entropy": 2.164503052830696, "epoch": 0.1377551020408163, "frac_reward_zero_std": 0.625, "grad_norm": 0.37109375, "learning_rate": 8.6281179138322e-07, "loss": 0.0063, "num_tokens": 23638156.0, "reward": 0.1484375, "reward_std": 0.47709178924560547, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 243, "step_time": 297.9009611783549 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3852.0, "completions/mean_length": 1256.703125, "completions/mean_terminated_length": 962.9827270507812, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "entropy": 3.2784062288701534, "epoch": 0.1383219954648526, "frac_reward_zero_std": 0.5, "grad_norm": 0.306640625, "learning_rate": 8.622448979591836e-07, "loss": 0.0578, "num_tokens": 23729785.0, "reward": 0.1796875, "reward_std": 0.5445777773857117, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 244, "step_time": 289.242567432113 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3455.0, "completions/mean_length": 1297.296875, "completions/mean_terminated_length": 716.4339599609375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.7673447020351887, "epoch": 0.1388888888888889, "frac_reward_zero_std": 0.875, "grad_norm": 0.1240234375, "learning_rate": 8.616780045351474e-07, "loss": 0.0088, "num_tokens": 23826884.0, "reward": 0.09375, "reward_std": 0.29378482699394226, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 245, "step_time": 316.948717080988 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3983.0, "completions/mean_length": 1402.921875, "completions/mean_terminated_length": 1018.1964721679688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.7002098821103573, "epoch": 0.13945578231292516, "frac_reward_zero_std": 0.75, "grad_norm": 0.21484375, "learning_rate": 8.611111111111111e-07, "loss": -0.0184, "num_tokens": 23930087.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 246, "step_time": 295.5066711353138 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4048.0, "completions/mean_length": 1391.359375, "completions/mean_terminated_length": 1059.2105712890625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.7934828735888004, "epoch": 0.14002267573696145, "frac_reward_zero_std": 0.625, "grad_norm": 0.224609375, "learning_rate": 8.605442176870747e-07, "loss": 0.039, "num_tokens": 24034950.0, "reward": 0.171875, "reward_std": 0.528690755367279, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 247, "step_time": 320.29575569648296 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4022.0, "completions/mean_length": 1085.359375, "completions/mean_terminated_length": 715.631591796875, "completions/min_length": 10.0, "completions/min_terminated_length": 10.0, "entropy": 2.4736474193632603, "epoch": 0.14058956916099774, "frac_reward_zero_std": 0.75, "grad_norm": 0.251953125, "learning_rate": 8.599773242630385e-07, "loss": 0.0231, "num_tokens": 24114565.0, "reward": 0.15625, "reward_std": 0.4443204402923584, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 248, "step_time": 292.1635997388512 }, { "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": 0.296875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3950.0, "completions/mean_length": 1851.671875, "completions/mean_terminated_length": 904.0667114257812, "completions/min_length": 126.0, "completions/min_terminated_length": 126.0, "entropy": 3.3864914886653423, "epoch": 0.141156462585034, "frac_reward_zero_std": 0.75, "grad_norm": 0.150390625, "learning_rate": 8.594104308390022e-07, "loss": 0.0031, "num_tokens": 24246208.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 249, "step_time": 297.76085374876857 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3766.0, "completions/mean_length": 1138.359375, "completions/mean_terminated_length": 775.140380859375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.169401206076145, "epoch": 0.1417233560090703, "frac_reward_zero_std": 0.875, "grad_norm": 0.20703125, "learning_rate": 8.58843537414966e-07, "loss": 0.0188, "num_tokens": 24330431.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 250, "step_time": 291.3457045610994 }, { "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": 0.25, "completions/max_length": 4096.0, "completions/max_terminated_length": 3965.0, "completions/mean_length": 1495.75, "completions/mean_terminated_length": 629.0, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 2.532334677875042, "epoch": 0.14229024943310659, "frac_reward_zero_std": 0.75, "grad_norm": 0.1865234375, "learning_rate": 8.582766439909297e-07, "loss": -0.0173, "num_tokens": 24438383.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 251, "step_time": 293.751147121191 }, { "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": 4096.0, "completions/max_terminated_length": 3714.0, "completions/mean_length": 936.046875, "completions/mean_terminated_length": 725.3833618164062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.8772873617708683, "epoch": 0.14285714285714285, "frac_reward_zero_std": 0.75, "grad_norm": 0.306640625, "learning_rate": 8.577097505668934e-07, "loss": 0.0716, "num_tokens": 24510306.0, "reward": 0.21875, "reward_std": 0.5900968313217163, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 252, "step_time": 289.76721812132746 }, { "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": 4096.0, "completions/max_terminated_length": 3876.0, "completions/mean_length": 946.625, "completions/mean_terminated_length": 736.6666870117188, "completions/min_length": 50.0, "completions/min_terminated_length": 50.0, "entropy": 2.8776904828846455, "epoch": 0.14342403628117914, "frac_reward_zero_std": 0.5, "grad_norm": 0.353515625, "learning_rate": 8.57142857142857e-07, "loss": 0.0881, "num_tokens": 24581746.0, "reward": 0.171875, "reward_std": 0.5434950590133667, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 253, "step_time": 288.82005726359785 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3496.0, "completions/mean_length": 1310.09375, "completions/mean_terminated_length": 912.107177734375, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "entropy": 4.113108057528734, "epoch": 0.14399092970521543, "frac_reward_zero_std": 0.875, "grad_norm": 0.1689453125, "learning_rate": 8.565759637188208e-07, "loss": 0.0293, "num_tokens": 24677800.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 254, "step_time": 293.571475405246 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3811.0, "completions/mean_length": 1182.3125, "completions/mean_terminated_length": 642.74072265625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.2622944563627243, "epoch": 0.1445578231292517, "frac_reward_zero_std": 0.75, "grad_norm": 0.2216796875, "learning_rate": 8.560090702947845e-07, "loss": -0.0382, "num_tokens": 24764820.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 255, "step_time": 289.96930027008057 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3109.0, "completions/mean_length": 1085.875, "completions/mean_terminated_length": 655.857177734375, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 2.153046991676092, "epoch": 0.14512471655328799, "frac_reward_zero_std": 0.875, "grad_norm": 0.2080078125, "learning_rate": 8.554421768707483e-07, "loss": 0.0088, "num_tokens": 24848284.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 256, "step_time": 310.2984448345378 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4015.0, "completions/mean_length": 1124.59375, "completions/mean_terminated_length": 817.2069091796875, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "entropy": 2.449272785335779, "epoch": 0.14569160997732428, "frac_reward_zero_std": 0.75, "grad_norm": 0.26171875, "learning_rate": 8.54875283446712e-07, "loss": 0.0355, "num_tokens": 24934018.0, "reward": 0.1484375, "reward_std": 0.5246290564537048, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 257, "step_time": 308.3471300471574 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3945.0, "completions/mean_length": 1355.796875, "completions/mean_terminated_length": 1019.2807006835938, "completions/min_length": 16.0, "completions/min_terminated_length": 16.0, "entropy": 3.478306507691741, "epoch": 0.14625850340136054, "frac_reward_zero_std": 0.75, "grad_norm": 0.2109375, "learning_rate": 8.543083900226758e-07, "loss": -0.0084, "num_tokens": 25031149.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 258, "step_time": 287.35603701416403 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3580.0, "completions/mean_length": 1468.40625, "completions/mean_terminated_length": 981.8148193359375, "completions/min_length": 236.0, "completions/min_terminated_length": 236.0, "entropy": 3.7615602426230907, "epoch": 0.14682539682539683, "frac_reward_zero_std": 0.875, "grad_norm": 0.1513671875, "learning_rate": 8.537414965986394e-07, "loss": 0.0076, "num_tokens": 25136863.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 259, "step_time": 297.84377315081656 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3869.0, "completions/mean_length": 1225.703125, "completions/mean_terminated_length": 815.6607666015625, "completions/min_length": 17.0, "completions/min_terminated_length": 17.0, "entropy": 3.4640529341995716, "epoch": 0.1473922902494331, "frac_reward_zero_std": 0.5, "grad_norm": 0.30859375, "learning_rate": 8.531746031746031e-07, "loss": -0.0409, "num_tokens": 25225876.0, "reward": 0.40625, "reward_std": 0.7811011672019958, "rewards/accuracy_reward_func/mean": 0.375, "rewards/accuracy_reward_func/std": 0.7867957949638367, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 260, "step_time": 292.2113784374669 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3578.0, "completions/mean_length": 1404.390625, "completions/mean_terminated_length": 783.25, "completions/min_length": 60.0, "completions/min_terminated_length": 60.0, "entropy": 2.998564811423421, "epoch": 0.14795918367346939, "frac_reward_zero_std": 0.625, "grad_norm": 0.2177734375, "learning_rate": 8.526077097505668e-07, "loss": 0.0184, "num_tokens": 25327797.0, "reward": 0.3203125, "reward_std": 0.6978833675384521, "rewards/accuracy_reward_func/mean": 0.28125, "rewards/accuracy_reward_func/std": 0.7007648944854736, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 261, "step_time": 297.0823359033093 }, { "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": 0.03125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4063.0, "completions/mean_length": 1167.328125, "completions/mean_terminated_length": 1072.8548583984375, "completions/min_length": 178.0, "completions/min_terminated_length": 178.0, "entropy": 3.6199896298348904, "epoch": 0.14852607709750568, "frac_reward_zero_std": 0.75, "grad_norm": 0.263671875, "learning_rate": 8.520408163265306e-07, "loss": 0.0266, "num_tokens": 25415178.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 262, "step_time": 297.0515417903662 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3808.0, "completions/mean_length": 1347.6875, "completions/mean_terminated_length": 1010.1754760742188, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "entropy": 3.2722493298351765, "epoch": 0.14909297052154194, "frac_reward_zero_std": 0.75, "grad_norm": 0.240234375, "learning_rate": 8.514739229024943e-07, "loss": -0.0253, "num_tokens": 25513334.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 263, "step_time": 289.5084929149598 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4089.0, "completions/mean_length": 1255.859375, "completions/mean_terminated_length": 850.1250610351562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.370054867118597, "epoch": 0.14965986394557823, "frac_reward_zero_std": 0.625, "grad_norm": 0.29296875, "learning_rate": 8.509070294784581e-07, "loss": -0.0059, "num_tokens": 25605181.0, "reward": 0.171875, "reward_std": 0.49776285886764526, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 264, "step_time": 290.37313639745116 }, { "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": 0.015625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3144.0, "completions/mean_length": 859.65625, "completions/mean_terminated_length": 808.2857666015625, "completions/min_length": 18.0, "completions/min_terminated_length": 18.0, "entropy": 3.8536425456404686, "epoch": 0.15022675736961452, "frac_reward_zero_std": 0.75, "grad_norm": 0.29296875, "learning_rate": 8.503401360544217e-07, "loss": 0.0229, "num_tokens": 25672167.0, "reward": 0.203125, "reward_std": 0.6023679971694946, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 265, "step_time": 294.0703383414075 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4009.0, "completions/mean_length": 1567.828125, "completions/mean_terminated_length": 1099.648193359375, "completions/min_length": 25.0, "completions/min_terminated_length": 25.0, "entropy": 5.198573417961597, "epoch": 0.15079365079365079, "frac_reward_zero_std": 0.875, "grad_norm": 0.134765625, "learning_rate": 8.497732426303855e-07, "loss": 0.0398, "num_tokens": 25783116.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 266, "step_time": 290.0501378560439 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3508.0, "completions/mean_length": 1046.40625, "completions/mean_terminated_length": 671.8947143554688, "completions/min_length": 35.0, "completions/min_terminated_length": 35.0, "entropy": 3.150088220834732, "epoch": 0.15136054421768708, "frac_reward_zero_std": 0.75, "grad_norm": 0.2490234375, "learning_rate": 8.492063492063492e-07, "loss": 0.0193, "num_tokens": 25862270.0, "reward": 0.1640625, "reward_std": 0.5274578928947449, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 267, "step_time": 295.3273321315646 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3815.0, "completions/mean_length": 979.671875, "completions/mean_terminated_length": 715.5762939453125, "completions/min_length": 68.0, "completions/min_terminated_length": 68.0, "entropy": 3.5900011137127876, "epoch": 0.15192743764172337, "frac_reward_zero_std": 0.75, "grad_norm": 0.26953125, "learning_rate": 8.486394557823128e-07, "loss": 0.03, "num_tokens": 25939529.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 268, "step_time": 305.9564151559025 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3940.0, "completions/mean_length": 893.734375, "completions/mean_terminated_length": 736.245849609375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 2.7484741285443306, "epoch": 0.15249433106575963, "frac_reward_zero_std": 0.5, "grad_norm": 0.42578125, "learning_rate": 8.480725623582766e-07, "loss": 0.0158, "num_tokens": 26009704.0, "reward": 0.2578125, "reward_std": 0.629891574382782, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 269, "step_time": 304.0618146182969 }, { "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": 4096.0, "completions/max_terminated_length": 3891.0, "completions/mean_length": 1002.703125, "completions/mean_terminated_length": 796.4833984375, "completions/min_length": 44.0, "completions/min_terminated_length": 44.0, "entropy": 3.7282408848404884, "epoch": 0.15306122448979592, "frac_reward_zero_std": 0.5, "grad_norm": 0.34765625, "learning_rate": 8.475056689342403e-07, "loss": 0.0477, "num_tokens": 26086149.0, "reward": 0.296875, "reward_std": 0.7544510364532471, "rewards/accuracy_reward_func/mean": 0.25, "rewards/accuracy_reward_func/std": 0.6666666865348816, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 270, "step_time": 300.7984333410859 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3989.0, "completions/mean_length": 1462.75, "completions/mean_terminated_length": 975.1111450195312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.990269973874092, "epoch": 0.1536281179138322, "frac_reward_zero_std": 0.75, "grad_norm": 0.203125, "learning_rate": 8.469387755102041e-07, "loss": 0.0466, "num_tokens": 26191237.0, "reward": 0.125, "reward_std": 0.4364357888698578, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 271, "step_time": 292.5327200219035 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3772.0, "completions/mean_length": 1295.578125, "completions/mean_terminated_length": 895.5178833007812, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 3.2695842906832695, "epoch": 0.15419501133786848, "frac_reward_zero_std": 0.75, "grad_norm": 0.2392578125, "learning_rate": 8.463718820861677e-07, "loss": 0.0, "num_tokens": 26286122.0, "reward": 0.0625, "reward_std": 0.1666666716337204, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 272, "step_time": 288.95768870413303 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4008.0, "completions/mean_length": 1152.921875, "completions/mean_terminated_length": 903.5084838867188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.692918371409178, "epoch": 0.15476190476190477, "frac_reward_zero_std": 0.5, "grad_norm": 0.376953125, "learning_rate": 8.458049886621315e-07, "loss": 0.0784, "num_tokens": 26371765.0, "reward": 0.140625, "reward_std": 0.4917473793029785, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 273, "step_time": 295.4737097416073 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3831.0, "completions/mean_length": 1018.234375, "completions/mean_terminated_length": 866.8687744140625, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 3.9792496412992477, "epoch": 0.15532879818594103, "frac_reward_zero_std": 0.625, "grad_norm": 0.375, "learning_rate": 8.452380952380952e-07, "loss": -0.0519, "num_tokens": 26448052.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 274, "step_time": 294.20816101972014 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3398.0, "completions/mean_length": 1288.484375, "completions/mean_terminated_length": 768.5740966796875, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "entropy": 3.9989641718566418, "epoch": 0.15589569160997732, "frac_reward_zero_std": 0.625, "grad_norm": 0.208984375, "learning_rate": 8.44671201814059e-07, "loss": 0.0087, "num_tokens": 26541003.0, "reward": 0.296875, "reward_std": 0.6884463429450989, "rewards/accuracy_reward_func/mean": 0.25, "rewards/accuracy_reward_func/std": 0.6666666865348816, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 275, "step_time": 289.7646218566224 }, { "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": 0.234375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4013.0, "completions/mean_length": 1927.5625, "completions/mean_terminated_length": 1263.755126953125, "completions/min_length": 129.0, "completions/min_terminated_length": 129.0, "entropy": 3.7165113035589457, "epoch": 0.1564625850340136, "frac_reward_zero_std": 0.75, "grad_norm": 0.15625, "learning_rate": 8.441043083900226e-07, "loss": -0.0189, "num_tokens": 26676503.0, "reward": 0.09375, "reward_std": 0.426084041595459, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 276, "step_time": 292.25411548279226 }, { "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": 0.03125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3567.0, "completions/mean_length": 753.5, "completions/mean_terminated_length": 645.6774291992188, "completions/min_length": 112.0, "completions/min_terminated_length": 112.0, "entropy": 2.645106505602598, "epoch": 0.15702947845804988, "frac_reward_zero_std": 0.625, "grad_norm": 0.388671875, "learning_rate": 8.435374149659864e-07, "loss": 0.0013, "num_tokens": 26737199.0, "reward": 0.2578125, "reward_std": 0.629891574382782, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 277, "step_time": 291.609153280966 }, { "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": 4096.0, "completions/max_terminated_length": 4033.0, "completions/mean_length": 1143.5, "completions/mean_terminated_length": 946.6666870117188, "completions/min_length": 161.0, "completions/min_terminated_length": 161.0, "entropy": 2.7190456688404083, "epoch": 0.15759637188208617, "frac_reward_zero_std": 0.875, "grad_norm": 0.123046875, "learning_rate": 8.4297052154195e-07, "loss": 0.0074, "num_tokens": 26823495.0, "reward": 0.1171875, "reward_std": 0.4519989788532257, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 278, "step_time": 301.25380596797913 }, { "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": 4096.0, "completions/max_terminated_length": 3969.0, "completions/mean_length": 1208.171875, "completions/mean_terminated_length": 1015.6500244140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.420939952135086, "epoch": 0.15816326530612246, "frac_reward_zero_std": 0.75, "grad_norm": 0.26953125, "learning_rate": 8.424036281179138e-07, "loss": -0.0159, "num_tokens": 26911730.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 279, "step_time": 292.37590639013797 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3969.0, "completions/mean_length": 1335.09375, "completions/mean_terminated_length": 883.30908203125, "completions/min_length": 25.0, "completions/min_terminated_length": 25.0, "entropy": 3.6540948897600174, "epoch": 0.15873015873015872, "frac_reward_zero_std": 0.875, "grad_norm": 0.1611328125, "learning_rate": 8.418367346938775e-07, "loss": -0.0055, "num_tokens": 27007784.0, "reward": 0.0703125, "reward_std": 0.33174845576286316, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 280, "step_time": 287.1056828228757 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3754.0, "completions/mean_length": 1120.515625, "completions/mean_terminated_length": 868.35595703125, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 3.532376177608967, "epoch": 0.159297052154195, "frac_reward_zero_std": 0.625, "grad_norm": 0.46875, "learning_rate": 8.412698412698413e-07, "loss": -0.1828, "num_tokens": 27091009.0, "reward": 0.125, "reward_std": 0.4714045226573944, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 281, "step_time": 295.76803965866566 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3972.0, "completions/mean_length": 1436.90625, "completions/mean_terminated_length": 944.4815063476562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.039565362036228, "epoch": 0.1598639455782313, "frac_reward_zero_std": 0.625, "grad_norm": 0.248046875, "learning_rate": 8.40702947845805e-07, "loss": 0.0313, "num_tokens": 27194355.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 282, "step_time": 289.60914684366435 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4048.0, "completions/mean_length": 1240.1875, "completions/mean_terminated_length": 832.2142944335938, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.066278997808695, "epoch": 0.16043083900226757, "frac_reward_zero_std": 0.875, "grad_norm": 0.09423828125, "learning_rate": 8.401360544217688e-07, "loss": 0.0053, "num_tokens": 27287391.0, "reward": 0.09375, "reward_std": 0.3435921370983124, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 283, "step_time": 299.21620115544647 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3833.0, "completions/mean_length": 1115.921875, "completions/mean_terminated_length": 863.3728637695312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.069147702306509, "epoch": 0.16099773242630386, "frac_reward_zero_std": 0.625, "grad_norm": 0.275390625, "learning_rate": 8.395691609977323e-07, "loss": -0.0164, "num_tokens": 27369802.0, "reward": 0.1484375, "reward_std": 0.49344661831855774, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 284, "step_time": 291.6855736114085 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3442.0, "completions/mean_length": 1112.84375, "completions/mean_terminated_length": 746.4912109375, "completions/min_length": 23.0, "completions/min_terminated_length": 23.0, "entropy": 3.147010438144207, "epoch": 0.16156462585034015, "frac_reward_zero_std": 0.625, "grad_norm": 0.3359375, "learning_rate": 8.39002267573696e-07, "loss": -0.0192, "num_tokens": 27452448.0, "reward": 0.1796875, "reward_std": 0.5445777773857117, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 285, "step_time": 295.8786328099668 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3757.0, "completions/mean_length": 1405.015625, "completions/mean_terminated_length": 1126.637939453125, "completions/min_length": 72.0, "completions/min_terminated_length": 72.0, "entropy": 5.521326437592506, "epoch": 0.1621315192743764, "frac_reward_zero_std": 0.75, "grad_norm": 0.185546875, "learning_rate": 8.384353741496598e-07, "loss": 0.0505, "num_tokens": 27554321.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 286, "step_time": 294.3407342750579 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3964.0, "completions/mean_length": 1361.078125, "completions/mean_terminated_length": 854.6111450195312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.5164739042520523, "epoch": 0.1626984126984127, "frac_reward_zero_std": 0.875, "grad_norm": 0.0732421875, "learning_rate": 8.378684807256235e-07, "loss": 0.0017, "num_tokens": 27653374.0, "reward": 0.1171875, "reward_std": 0.4519989788532257, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 287, "step_time": 293.39336588140577 }, { "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": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 2766.0, "completions/mean_length": 1531.0, "completions/mean_terminated_length": 812.7999877929688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.3258663713932037, "epoch": 0.16326530612244897, "frac_reward_zero_std": 0.875, "grad_norm": 0.1416015625, "learning_rate": 8.373015873015873e-07, "loss": -0.01, "num_tokens": 27764206.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 288, "step_time": 293.5519551774487 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3920.0, "completions/mean_length": 1234.671875, "completions/mean_terminated_length": 825.9107666015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.563442576676607, "epoch": 0.16383219954648526, "frac_reward_zero_std": 0.875, "grad_norm": 0.16796875, "learning_rate": 8.36734693877551e-07, "loss": 0.0538, "num_tokens": 27855057.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 289, "step_time": 290.0854655979201 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3750.0, "completions/mean_length": 1312.671875, "completions/mean_terminated_length": 970.8596801757812, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 4.127264823764563, "epoch": 0.16439909297052155, "frac_reward_zero_std": 0.625, "grad_norm": 0.294921875, "learning_rate": 8.361678004535147e-07, "loss": -0.0188, "num_tokens": 27950628.0, "reward": 0.0703125, "reward_std": 0.35486623644828796, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 290, "step_time": 291.64527351316065 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3994.0, "completions/mean_length": 1398.359375, "completions/mean_terminated_length": 1012.982177734375, "completions/min_length": 24.0, "completions/min_terminated_length": 24.0, "entropy": 3.6126014180481434, "epoch": 0.1649659863945578, "frac_reward_zero_std": 0.75, "grad_norm": 0.26171875, "learning_rate": 8.356009070294783e-07, "loss": -0.069, "num_tokens": 28053083.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 291, "step_time": 304.7845337698236 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3263.0, "completions/mean_length": 1011.578125, "completions/mean_terminated_length": 750.1864624023438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.8906866386532784, "epoch": 0.1655328798185941, "frac_reward_zero_std": 0.75, "grad_norm": 0.2734375, "learning_rate": 8.350340136054421e-07, "loss": 0.0055, "num_tokens": 28133432.0, "reward": 0.1328125, "reward_std": 0.5058560967445374, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 292, "step_time": 308.4352606162429 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3983.0, "completions/mean_length": 1136.640625, "completions/mean_terminated_length": 830.5, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 3.1073951721191406, "epoch": 0.1660997732426304, "frac_reward_zero_std": 0.75, "grad_norm": 0.2099609375, "learning_rate": 8.344671201814058e-07, "loss": 0.0566, "num_tokens": 28218497.0, "reward": 0.1015625, "reward_std": 0.40999117493629456, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 293, "step_time": 297.7840570639819 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4085.0, "completions/mean_length": 1029.5, "completions/mean_terminated_length": 878.6884765625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.2432908341288567, "epoch": 0.16666666666666666, "frac_reward_zero_std": 0.625, "grad_norm": 0.28125, "learning_rate": 8.339002267573696e-07, "loss": 0.0124, "num_tokens": 28298473.0, "reward": 0.25, "reward_std": 0.6546536684036255, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 294, "step_time": 300.63302401360124 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3131.0, "completions/mean_length": 929.875, "completions/mean_terminated_length": 602.3448486328125, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 2.910718649625778, "epoch": 0.16723356009070295, "frac_reward_zero_std": 0.75, "grad_norm": 0.205078125, "learning_rate": 8.333333333333333e-07, "loss": 0.0047, "num_tokens": 28370089.0, "reward": 0.078125, "reward_std": 0.2847827076911926, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 295, "step_time": 295.5374428220093 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4079.0, "completions/mean_length": 1041.5625, "completions/mean_terminated_length": 891.34423828125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.183424171060324, "epoch": 0.16780045351473924, "frac_reward_zero_std": 0.75, "grad_norm": 0.275390625, "learning_rate": 8.327664399092971e-07, "loss": -0.0004, "num_tokens": 28448221.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 296, "step_time": 291.7701356653124 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4023.0, "completions/mean_length": 1469.453125, "completions/mean_terminated_length": 983.0555419921875, "completions/min_length": 38.0, "completions/min_terminated_length": 38.0, "entropy": 3.3522771634161472, "epoch": 0.1683673469387755, "frac_reward_zero_std": 0.875, "grad_norm": 0.1572265625, "learning_rate": 8.321995464852607e-07, "loss": 0.0624, "num_tokens": 28554090.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 297, "step_time": 290.37179633043706 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4090.0, "completions/mean_length": 1065.140625, "completions/mean_terminated_length": 916.0819091796875, "completions/min_length": 122.0, "completions/min_terminated_length": 122.0, "entropy": 3.6264598928391933, "epoch": 0.1689342403628118, "frac_reward_zero_std": 0.875, "grad_norm": 0.197265625, "learning_rate": 8.316326530612245e-07, "loss": -0.0006, "num_tokens": 28633771.0, "reward": 0.03125, "reward_std": 0.12198750674724579, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 298, "step_time": 289.74198976531625 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3413.0, "completions/mean_length": 1257.59375, "completions/mean_terminated_length": 731.9629516601562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.8235544338822365, "epoch": 0.16950113378684808, "frac_reward_zero_std": 0.875, "grad_norm": 0.2138671875, "learning_rate": 8.310657596371881e-07, "loss": 0.0584, "num_tokens": 28726809.0, "reward": 0.046875, "reward_std": 0.14689241349697113, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 299, "step_time": 298.02177797444165 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3896.0, "completions/mean_length": 1213.59375, "completions/mean_terminated_length": 859.614013671875, "completions/min_length": 79.0, "completions/min_terminated_length": 79.0, "entropy": 3.136536069214344, "epoch": 0.17006802721088435, "frac_reward_zero_std": 0.875, "grad_norm": 0.201171875, "learning_rate": 8.304988662131519e-07, "loss": 0.0282, "num_tokens": 28817535.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 300, "step_time": 298.4104125360027 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3977.0, "completions/mean_length": 1086.96875, "completions/mean_terminated_length": 938.9835815429688, "completions/min_length": 69.0, "completions/min_terminated_length": 69.0, "entropy": 4.44902740418911, "epoch": 0.17063492063492064, "frac_reward_zero_std": 0.625, "grad_norm": 0.2890625, "learning_rate": 8.299319727891156e-07, "loss": 0.0085, "num_tokens": 28899477.0, "reward": 0.1484375, "reward_std": 0.47709178924560547, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 301, "step_time": 293.0110935186967 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3765.0, "completions/mean_length": 1278.453125, "completions/mean_terminated_length": 875.9464721679688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.5456941816955805, "epoch": 0.1712018140589569, "frac_reward_zero_std": 0.75, "grad_norm": 0.25, "learning_rate": 8.293650793650794e-07, "loss": -0.0173, "num_tokens": 28993426.0, "reward": 0.1953125, "reward_std": 0.6016985774040222, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 302, "step_time": 290.9091587802395 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3951.0, "completions/mean_length": 1041.546875, "completions/mean_terminated_length": 891.3278198242188, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "entropy": 3.8210051991045475, "epoch": 0.1717687074829932, "frac_reward_zero_std": 0.625, "grad_norm": 0.31640625, "learning_rate": 8.28798185941043e-07, "loss": 0.0747, "num_tokens": 29071709.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 303, "step_time": 291.6461437251419 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3922.0, "completions/mean_length": 1016.0, "completions/mean_terminated_length": 754.9830322265625, "completions/min_length": 83.0, "completions/min_terminated_length": 83.0, "entropy": 3.1986773125827312, "epoch": 0.17233560090702948, "frac_reward_zero_std": 0.75, "grad_norm": 0.2470703125, "learning_rate": 8.282312925170068e-07, "loss": 0.0172, "num_tokens": 29146213.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 304, "step_time": 286.11513439565897 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3994.0, "completions/mean_length": 1154.609375, "completions/mean_terminated_length": 850.3275756835938, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "entropy": 4.014871634542942, "epoch": 0.17290249433106575, "frac_reward_zero_std": 0.625, "grad_norm": 0.369140625, "learning_rate": 8.276643990929705e-07, "loss": -0.052, "num_tokens": 29232268.0, "reward": 0.1015625, "reward_std": 0.3692515790462494, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 305, "step_time": 301.72234365716577 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3917.0, "completions/mean_length": 1428.734375, "completions/mean_terminated_length": 992.272705078125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.376392114907503, "epoch": 0.17346938775510204, "frac_reward_zero_std": 0.625, "grad_norm": 0.1845703125, "learning_rate": 8.270975056689343e-07, "loss": 0.0337, "num_tokens": 29337259.0, "reward": 0.2734375, "reward_std": 0.7449815273284912, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 306, "step_time": 300.49586669448763 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4075.0, "completions/mean_length": 1308.984375, "completions/mean_terminated_length": 1020.6724243164062, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.3659167028963566, "epoch": 0.17403628117913833, "frac_reward_zero_std": 0.5, "grad_norm": 0.353515625, "learning_rate": 8.26530612244898e-07, "loss": 0.0935, "num_tokens": 29432562.0, "reward": 0.1953125, "reward_std": 0.5463964939117432, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 307, "step_time": 297.3954945784062 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3754.0, "completions/mean_length": 1249.78125, "completions/mean_terminated_length": 843.1785888671875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.9153324998915195, "epoch": 0.1746031746031746, "frac_reward_zero_std": 0.625, "grad_norm": 0.234375, "learning_rate": 8.259637188208616e-07, "loss": 0.0041, "num_tokens": 29530020.0, "reward": 0.28125, "reward_std": 0.7007648944854736, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 308, "step_time": 322.4700731197372 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4046.0, "completions/mean_length": 1069.484375, "completions/mean_terminated_length": 813.0, "completions/min_length": 141.0, "completions/min_terminated_length": 141.0, "entropy": 3.3399731442332268, "epoch": 0.17517006802721088, "frac_reward_zero_std": 0.75, "grad_norm": 0.25390625, "learning_rate": 8.253968253968253e-07, "loss": 0.0272, "num_tokens": 29610987.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 309, "step_time": 291.58139999583364 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3754.0, "completions/mean_length": 1078.03125, "completions/mean_terminated_length": 765.8275756835938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.948049958795309, "epoch": 0.17573696145124718, "frac_reward_zero_std": 0.75, "grad_norm": 0.201171875, "learning_rate": 8.24829931972789e-07, "loss": 0.019, "num_tokens": 29691661.0, "reward": 0.1875, "reward_std": 0.5, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 310, "step_time": 293.682350859046 }, { "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": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3575.0, "completions/mean_length": 1424.96875, "completions/mean_terminated_length": 744.11767578125, "completions/min_length": 26.0, "completions/min_terminated_length": 26.0, "entropy": 2.5001518316566944, "epoch": 0.17630385487528344, "frac_reward_zero_std": 0.75, "grad_norm": 0.181640625, "learning_rate": 8.242630385487528e-07, "loss": -0.0158, "num_tokens": 29794195.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 311, "step_time": 290.9478914504871 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3947.0, "completions/mean_length": 1281.71875, "completions/mean_terminated_length": 821.199951171875, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.540242325514555, "epoch": 0.17687074829931973, "frac_reward_zero_std": 0.5, "grad_norm": 0.275390625, "learning_rate": 8.236961451247165e-07, "loss": 0.0548, "num_tokens": 29887745.0, "reward": 0.2890625, "reward_std": 0.7003666162490845, "rewards/accuracy_reward_func/mean": 0.28125, "rewards/accuracy_reward_func/std": 0.7007648944854736, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 312, "step_time": 293.616348778829 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3581.0, "completions/mean_length": 962.25, "completions/mean_terminated_length": 808.131103515625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.187012553215027, "epoch": 0.17743764172335602, "frac_reward_zero_std": 0.5, "grad_norm": 0.28515625, "learning_rate": 8.231292517006803e-07, "loss": 0.0173, "num_tokens": 29960497.0, "reward": 0.15625, "reward_std": 0.49501484632492065, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 313, "step_time": 288.6750570386648 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3996.0, "completions/mean_length": 1446.0, "completions/mean_terminated_length": 955.25927734375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.385607846081257, "epoch": 0.17800453514739228, "frac_reward_zero_std": 0.875, "grad_norm": 0.1787109375, "learning_rate": 8.22562358276644e-07, "loss": 0.022, "num_tokens": 30064113.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 314, "step_time": 288.21967145614326 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3718.0, "completions/mean_length": 1296.296875, "completions/mean_terminated_length": 715.2264404296875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.9967053532600403, "epoch": 0.17857142857142858, "frac_reward_zero_std": 0.75, "grad_norm": 0.205078125, "learning_rate": 8.219954648526076e-07, "loss": 0.0107, "num_tokens": 30159260.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 315, "step_time": 301.35519457794726 }, { "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": 4096.0, "completions/max_terminated_length": 3953.0, "completions/mean_length": 1092.03125, "completions/mean_terminated_length": 891.7667236328125, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.5508852675557137, "epoch": 0.17913832199546487, "frac_reward_zero_std": 0.875, "grad_norm": 0.25390625, "learning_rate": 8.214285714285713e-07, "loss": 0.0176, "num_tokens": 30242230.0, "reward": 0.046875, "reward_std": 0.14689241349697113, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 316, "step_time": 299.0346463257447 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3781.0, "completions/mean_length": 1171.609375, "completions/mean_terminated_length": 869.086181640625, "completions/min_length": 14.0, "completions/min_terminated_length": 14.0, "entropy": 3.63652553409338, "epoch": 0.17970521541950113, "frac_reward_zero_std": 0.5, "grad_norm": 0.359375, "learning_rate": 8.208616780045351e-07, "loss": 0.0816, "num_tokens": 30329205.0, "reward": 0.1171875, "reward_std": 0.43408530950546265, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 317, "step_time": 292.3733606124297 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 1257.0, "completions/mean_length": 762.453125, "completions/mean_terminated_length": 417.60345458984375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 1.2139202319085598, "epoch": 0.18027210884353742, "frac_reward_zero_std": 0.75, "grad_norm": 0.3203125, "learning_rate": 8.202947845804988e-07, "loss": 0.0101, "num_tokens": 30390298.0, "reward": 0.1015625, "reward_std": 0.3692515790462494, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 318, "step_time": 297.21323311980814 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3501.0, "completions/mean_length": 1052.65625, "completions/mean_terminated_length": 678.9122924804688, "completions/min_length": 158.0, "completions/min_terminated_length": 158.0, "entropy": 2.5844947583973408, "epoch": 0.18083900226757368, "frac_reward_zero_std": 0.625, "grad_norm": 0.328125, "learning_rate": 8.197278911564626e-07, "loss": -0.0033, "num_tokens": 30468452.0, "reward": 0.1015625, "reward_std": 0.3692515790462494, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 319, "step_time": 287.7088434314355 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3865.0, "completions/mean_length": 1342.53125, "completions/mean_terminated_length": 832.629638671875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.4571900218725204, "epoch": 0.18140589569160998, "frac_reward_zero_std": 0.875, "grad_norm": 0.1416015625, "learning_rate": 8.191609977324263e-07, "loss": -0.0004, "num_tokens": 30565646.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 320, "step_time": 289.6377345332876 }, { "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": 4096.0, "completions/max_terminated_length": 3584.0, "completions/mean_length": 999.375, "completions/mean_terminated_length": 792.933349609375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.4800062365829945, "epoch": 0.18197278911564627, "frac_reward_zero_std": 0.625, "grad_norm": 0.322265625, "learning_rate": 8.1859410430839e-07, "loss": 0.0531, "num_tokens": 30641374.0, "reward": 0.25, "reward_std": 0.6299408078193665, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 321, "step_time": 292.96126566361636 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3794.0, "completions/mean_length": 1082.109375, "completions/mean_terminated_length": 711.9824829101562, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.381516419351101, "epoch": 0.18253968253968253, "frac_reward_zero_std": 0.625, "grad_norm": 0.279296875, "learning_rate": 8.180272108843537e-07, "loss": 0.0632, "num_tokens": 30722573.0, "reward": 0.171875, "reward_std": 0.5434950590133667, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 322, "step_time": 296.53174413740635 }, { "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": 0.21875, "completions/max_length": 4096.0, "completions/max_terminated_length": 2639.0, "completions/mean_length": 1468.796875, "completions/mean_terminated_length": 733.1799926757812, "completions/min_length": 65.0, "completions/min_terminated_length": 65.0, "entropy": 2.0585499219596386, "epoch": 0.18310657596371882, "frac_reward_zero_std": 0.625, "grad_norm": 0.26953125, "learning_rate": 8.174603174603174e-07, "loss": -0.0296, "num_tokens": 30827456.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 323, "step_time": 289.30965161230415 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3585.0, "completions/mean_length": 1254.484375, "completions/mean_terminated_length": 905.5263061523438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.5128559842705727, "epoch": 0.1836734693877551, "frac_reward_zero_std": 0.75, "grad_norm": 0.224609375, "learning_rate": 8.168934240362811e-07, "loss": -0.0116, "num_tokens": 30918455.0, "reward": 0.109375, "reward_std": 0.4315783977508545, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 324, "step_time": 289.2214300278574 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4009.0, "completions/mean_length": 1172.1875, "completions/mean_terminated_length": 1028.3934326171875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.520883943885565, "epoch": 0.18424036281179137, "frac_reward_zero_std": 0.625, "grad_norm": 0.30078125, "learning_rate": 8.163265306122449e-07, "loss": 0.0129, "num_tokens": 31005915.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 325, "step_time": 295.2983176205307 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3725.0, "completions/mean_length": 1161.890625, "completions/mean_terminated_length": 858.362060546875, "completions/min_length": 109.0, "completions/min_terminated_length": 109.0, "entropy": 2.8527053482830524, "epoch": 0.18480725623582767, "frac_reward_zero_std": 0.625, "grad_norm": 0.345703125, "learning_rate": 8.157596371882086e-07, "loss": -0.0646, "num_tokens": 31092388.0, "reward": 0.15625, "reward_std": 0.49501484632492065, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 326, "step_time": 292.3190690325573 }, { "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": 4096.0, "completions/max_terminated_length": 3612.0, "completions/mean_length": 1134.671875, "completions/mean_terminated_length": 937.2500610351562, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.1747280284762383, "epoch": 0.18537414965986396, "frac_reward_zero_std": 0.625, "grad_norm": 0.3046875, "learning_rate": 8.151927437641724e-07, "loss": 0.0846, "num_tokens": 31177767.0, "reward": 0.171875, "reward_std": 0.49776285886764526, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 327, "step_time": 301.934424594976 }, { "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": 0.25, "completions/max_length": 4096.0, "completions/max_terminated_length": 4024.0, "completions/mean_length": 1899.515625, "completions/mean_terminated_length": 1167.354248046875, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.8632242269814014, "epoch": 0.18594104308390022, "frac_reward_zero_std": 0.75, "grad_norm": 0.17578125, "learning_rate": 8.14625850340136e-07, "loss": 0.0217, "num_tokens": 31310560.0, "reward": 0.1015625, "reward_std": 0.3692515790462494, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 328, "step_time": 288.49832714069635 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3925.0, "completions/mean_length": 1529.078125, "completions/mean_terminated_length": 936.7115478515625, "completions/min_length": 142.0, "completions/min_terminated_length": 142.0, "entropy": 3.7122786603868008, "epoch": 0.1865079365079365, "frac_reward_zero_std": 0.75, "grad_norm": 0.236328125, "learning_rate": 8.140589569160997e-07, "loss": -0.0115, "num_tokens": 31420869.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 329, "step_time": 292.37152093090117 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3782.0, "completions/mean_length": 1069.609375, "completions/mean_terminated_length": 756.5344848632812, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 2.990474697202444, "epoch": 0.1870748299319728, "frac_reward_zero_std": 0.75, "grad_norm": 0.232421875, "learning_rate": 8.134920634920635e-07, "loss": 0.0011, "num_tokens": 31500516.0, "reward": 0.1328125, "reward_std": 0.4386323094367981, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 330, "step_time": 289.3762530880049 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4020.0, "completions/mean_length": 1154.5625, "completions/mean_terminated_length": 734.357177734375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 2.997950706630945, "epoch": 0.18764172335600907, "frac_reward_zero_std": 0.875, "grad_norm": 0.1953125, "learning_rate": 8.129251700680271e-07, "loss": -0.0001, "num_tokens": 31585248.0, "reward": 0.015625, "reward_std": 0.08768405020236969, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 331, "step_time": 288.0672324942425 }, { "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": 4096.0, "completions/max_terminated_length": 3675.0, "completions/mean_length": 1097.765625, "completions/mean_terminated_length": 897.8833618164062, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 4.069846838712692, "epoch": 0.18820861678004536, "frac_reward_zero_std": 0.75, "grad_norm": 0.3203125, "learning_rate": 8.123582766439909e-07, "loss": 0.0991, "num_tokens": 31669993.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 332, "step_time": 303.92201447300613 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3817.0, "completions/mean_length": 1341.875, "completions/mean_terminated_length": 1003.6491088867188, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.1740684993565083, "epoch": 0.18877551020408162, "frac_reward_zero_std": 0.75, "grad_norm": 0.2099609375, "learning_rate": 8.117913832199546e-07, "loss": 0.0217, "num_tokens": 31767417.0, "reward": 0.1328125, "reward_std": 0.489915668964386, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 333, "step_time": 292.0739118931815 }, { "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": 0.015625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3941.0, "completions/mean_length": 843.875, "completions/mean_terminated_length": 792.2540283203125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.6779583133757114, "epoch": 0.1893424036281179, "frac_reward_zero_std": 0.75, "grad_norm": 0.376953125, "learning_rate": 8.112244897959183e-07, "loss": -0.0109, "num_tokens": 31832489.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 334, "step_time": 289.7429932355881 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3631.0, "completions/mean_length": 1274.1875, "completions/mean_terminated_length": 871.0714721679688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.734599068760872, "epoch": 0.1899092970521542, "frac_reward_zero_std": 0.75, "grad_norm": 0.2255859375, "learning_rate": 8.10657596371882e-07, "loss": -0.003, "num_tokens": 31926069.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 335, "step_time": 289.7558412672952 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3895.0, "completions/mean_length": 1368.453125, "completions/mean_terminated_length": 922.1272583007812, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.290547788143158, "epoch": 0.19047619047619047, "frac_reward_zero_std": 0.75, "grad_norm": 0.1962890625, "learning_rate": 8.100907029478458e-07, "loss": -0.0051, "num_tokens": 32024626.0, "reward": 0.109375, "reward_std": 0.4315783977508545, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 336, "step_time": 290.3555962955579 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3994.0, "completions/mean_length": 1159.625, "completions/mean_terminated_length": 855.862060546875, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.410203080624342, "epoch": 0.19104308390022676, "frac_reward_zero_std": 0.875, "grad_norm": 0.15625, "learning_rate": 8.095238095238095e-07, "loss": -0.0024, "num_tokens": 32110570.0, "reward": 0.0390625, "reward_std": 0.13524484634399414, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 337, "step_time": 296.57070900686085 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3897.0, "completions/mean_length": 1003.640625, "completions/mean_terminated_length": 683.7413940429688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.98492569103837, "epoch": 0.19160997732426305, "frac_reward_zero_std": 0.75, "grad_norm": 0.2451171875, "learning_rate": 8.089569160997733e-07, "loss": -0.0003, "num_tokens": 32188283.0, "reward": 0.1484375, "reward_std": 0.44257262349128723, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 338, "step_time": 294.97411496564746 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4028.0, "completions/mean_length": 1315.421875, "completions/mean_terminated_length": 918.1964721679688, "completions/min_length": 169.0, "completions/min_terminated_length": 169.0, "entropy": 3.99150575324893, "epoch": 0.1921768707482993, "frac_reward_zero_std": 0.625, "grad_norm": 0.27734375, "learning_rate": 8.083900226757369e-07, "loss": -0.0336, "num_tokens": 32285366.0, "reward": 0.1015625, "reward_std": 0.3692515790462494, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 339, "step_time": 302.7535767983645 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4093.0, "completions/mean_length": 1326.40625, "completions/mean_terminated_length": 1039.8966064453125, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 4.203001990914345, "epoch": 0.1927437641723356, "frac_reward_zero_std": 0.875, "grad_norm": 0.11474609375, "learning_rate": 8.078231292517006e-07, "loss": 0.0007, "num_tokens": 32381936.0, "reward": 0.0703125, "reward_std": 0.33174845576286316, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 340, "step_time": 291.1374069619924 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4011.0, "completions/mean_length": 1400.5625, "completions/mean_terminated_length": 1172.1356201171875, "completions/min_length": 77.0, "completions/min_terminated_length": 77.0, "entropy": 4.108578380197287, "epoch": 0.1933106575963719, "frac_reward_zero_std": 0.75, "grad_norm": 0.21875, "learning_rate": 8.072562358276643e-07, "loss": 0.0643, "num_tokens": 32485548.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 341, "step_time": 296.7473053280264 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3645.0, "completions/mean_length": 1393.078125, "completions/mean_terminated_length": 892.5370483398438, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 2.918832741677761, "epoch": 0.19387755102040816, "frac_reward_zero_std": 1.0, "grad_norm": 0.0, "learning_rate": 8.066893424036281e-07, "loss": 0.0, "num_tokens": 32585793.0, "reward": 0.0, "reward_std": 0.0, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 342, "step_time": 292.41653973422945 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3845.0, "completions/mean_length": 1193.40625, "completions/mean_terminated_length": 893.137939453125, "completions/min_length": 205.0, "completions/min_terminated_length": 205.0, "entropy": 2.4753906466066837, "epoch": 0.19444444444444445, "frac_reward_zero_std": 0.75, "grad_norm": 0.30859375, "learning_rate": 8.061224489795918e-07, "loss": 0.0039, "num_tokens": 32674323.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 343, "step_time": 293.4757890962064 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3682.0, "completions/mean_length": 1056.328125, "completions/mean_terminated_length": 798.7288208007812, "completions/min_length": 20.0, "completions/min_terminated_length": 20.0, "entropy": 2.791908524930477, "epoch": 0.19501133786848074, "frac_reward_zero_std": 0.75, "grad_norm": 0.234375, "learning_rate": 8.055555555555556e-07, "loss": 0.0158, "num_tokens": 32753120.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 344, "step_time": 293.0030291257426 }, { "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": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4002.0, "completions/mean_length": 1553.390625, "completions/mean_terminated_length": 905.2745361328125, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.106761857867241, "epoch": 0.195578231292517, "frac_reward_zero_std": 0.75, "grad_norm": 0.2197265625, "learning_rate": 8.049886621315193e-07, "loss": 0.0236, "num_tokens": 32866649.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 345, "step_time": 298.3969454225153 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3996.0, "completions/mean_length": 1171.875, "completions/mean_terminated_length": 693.3817749023438, "completions/min_length": 173.0, "completions/min_terminated_length": 173.0, "entropy": 2.4435928352177143, "epoch": 0.1961451247165533, "frac_reward_zero_std": 0.5, "grad_norm": 0.359375, "learning_rate": 8.04421768707483e-07, "loss": -0.0276, "num_tokens": 32954065.0, "reward": 0.265625, "reward_std": 0.6297439336776733, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 346, "step_time": 294.1192708387971 }, { "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": 4096.0, "completions/max_terminated_length": 3540.0, "completions/mean_length": 1185.8125, "completions/mean_terminated_length": 991.800048828125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 2.965976595878601, "epoch": 0.19671201814058956, "frac_reward_zero_std": 0.625, "grad_norm": 0.2734375, "learning_rate": 8.038548752834466e-07, "loss": 0.0068, "num_tokens": 33039957.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 347, "step_time": 287.24837697297335 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3757.0, "completions/mean_length": 1442.703125, "completions/mean_terminated_length": 1063.6607666015625, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.8474852554500103, "epoch": 0.19727891156462585, "frac_reward_zero_std": 0.75, "grad_norm": 0.2099609375, "learning_rate": 8.032879818594103e-07, "loss": 0.0384, "num_tokens": 33143306.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 348, "step_time": 290.6687171831727 }, { "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": 4096.0, "completions/max_terminated_length": 3542.0, "completions/mean_length": 1114.453125, "completions/mean_terminated_length": 915.6834106445312, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.210033155977726, "epoch": 0.19784580498866214, "frac_reward_zero_std": 0.625, "grad_norm": 0.33984375, "learning_rate": 8.027210884353741e-07, "loss": -0.0325, "num_tokens": 33227207.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 349, "step_time": 299.9674756946042 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3699.0, "completions/mean_length": 1139.15625, "completions/mean_terminated_length": 888.5762939453125, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "entropy": 2.9709309451282024, "epoch": 0.1984126984126984, "frac_reward_zero_std": 0.75, "grad_norm": 0.1904296875, "learning_rate": 8.021541950113378e-07, "loss": 0.0131, "num_tokens": 33316969.0, "reward": 0.1171875, "reward_std": 0.41539984941482544, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 350, "step_time": 340.1948959333822 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3753.0, "completions/mean_length": 865.5625, "completions/mean_terminated_length": 591.796630859375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.754661999642849, "epoch": 0.1989795918367347, "frac_reward_zero_std": 0.625, "grad_norm": 0.478515625, "learning_rate": 8.015873015873016e-07, "loss": 0.0932, "num_tokens": 33384901.0, "reward": 0.2421875, "reward_std": 0.5908843874931335, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 351, "step_time": 306.4861232312396 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4051.0, "completions/mean_length": 1256.140625, "completions/mean_terminated_length": 850.4464721679688, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "entropy": 3.122956484556198, "epoch": 0.19954648526077098, "frac_reward_zero_std": 0.5, "grad_norm": 0.283203125, "learning_rate": 8.010204081632652e-07, "loss": 0.0551, "num_tokens": 33476318.0, "reward": 0.2578125, "reward_std": 0.666620135307312, "rewards/accuracy_reward_func/mean": 0.25, "rewards/accuracy_reward_func/std": 0.6666666865348816, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 352, "step_time": 291.8524446422234 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3672.0, "completions/mean_length": 1041.34375, "completions/mean_terminated_length": 891.1146850585938, "completions/min_length": 48.0, "completions/min_terminated_length": 48.0, "entropy": 4.500285293906927, "epoch": 0.20011337868480725, "frac_reward_zero_std": 0.875, "grad_norm": 0.2138671875, "learning_rate": 8.00453514739229e-07, "loss": 0.0251, "num_tokens": 33554316.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 353, "step_time": 297.3388783968985 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3766.0, "completions/mean_length": 1027.71875, "completions/mean_terminated_length": 710.3103637695312, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.6725300680845976, "epoch": 0.20068027210884354, "frac_reward_zero_std": 0.75, "grad_norm": 0.2197265625, "learning_rate": 7.998866213151927e-07, "loss": 0.014, "num_tokens": 33631402.0, "reward": 0.1484375, "reward_std": 0.5092765688896179, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 354, "step_time": 293.5111241219565 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3766.0, "completions/mean_length": 1036.671875, "completions/mean_terminated_length": 720.1896362304688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.5119920670986176, "epoch": 0.20124716553287983, "frac_reward_zero_std": 0.75, "grad_norm": 0.2734375, "learning_rate": 7.993197278911564e-07, "loss": 0.017, "num_tokens": 33708029.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 355, "step_time": 288.01610595174134 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3636.0, "completions/mean_length": 1156.578125, "completions/mean_terminated_length": 795.5964965820312, "completions/min_length": 15.0, "completions/min_terminated_length": 15.0, "entropy": 2.926870696246624, "epoch": 0.2018140589569161, "frac_reward_zero_std": 0.625, "grad_norm": 0.341796875, "learning_rate": 7.987528344671201e-07, "loss": -0.0853, "num_tokens": 33797706.0, "reward": 0.15625, "reward_std": 0.49501484632492065, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 356, "step_time": 316.6249850699678 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3899.0, "completions/mean_length": 1299.0625, "completions/mean_terminated_length": 955.5789794921875, "completions/min_length": 43.0, "completions/min_terminated_length": 43.0, "entropy": 3.0896060839295387, "epoch": 0.20238095238095238, "frac_reward_zero_std": 0.375, "grad_norm": 0.361328125, "learning_rate": 7.981859410430839e-07, "loss": 0.1148, "num_tokens": 33891158.0, "reward": 0.296875, "reward_std": 0.6649904847145081, "rewards/accuracy_reward_func/mean": 0.25, "rewards/accuracy_reward_func/std": 0.6666666865348816, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 357, "step_time": 291.90663736127317 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3048.0, "completions/mean_length": 1077.046875, "completions/mean_terminated_length": 706.2982788085938, "completions/min_length": 73.0, "completions/min_terminated_length": 73.0, "entropy": 1.8321760259568691, "epoch": 0.20294784580498867, "frac_reward_zero_std": 0.625, "grad_norm": 0.333984375, "learning_rate": 7.976190476190476e-07, "loss": 0.0227, "num_tokens": 33972569.0, "reward": 0.15625, "reward_std": 0.49501484632492065, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 358, "step_time": 304.3349149832502 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4076.0, "completions/mean_length": 1328.984375, "completions/mean_terminated_length": 989.1754760742188, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.4920119643211365, "epoch": 0.20351473922902494, "frac_reward_zero_std": 0.5, "grad_norm": 0.341796875, "learning_rate": 7.970521541950113e-07, "loss": -0.0037, "num_tokens": 34069616.0, "reward": 0.15625, "reward_std": 0.49501484632492065, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 359, "step_time": 297.3019419154152 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4089.0, "completions/mean_length": 1408.34375, "completions/mean_terminated_length": 910.629638671875, "completions/min_length": 50.0, "completions/min_terminated_length": 50.0, "entropy": 3.2245482206344604, "epoch": 0.20408163265306123, "frac_reward_zero_std": 0.5, "grad_norm": 0.279296875, "learning_rate": 7.96485260770975e-07, "loss": 0.0111, "num_tokens": 34170438.0, "reward": 0.171875, "reward_std": 0.5434950590133667, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 360, "step_time": 293.22710838448256 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4014.0, "completions/mean_length": 1489.453125, "completions/mean_terminated_length": 1006.75927734375, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 3.7275636307895184, "epoch": 0.2046485260770975, "frac_reward_zero_std": 0.875, "grad_norm": 0.10693359375, "learning_rate": 7.959183673469388e-07, "loss": 0.0095, "num_tokens": 34278091.0, "reward": 0.09375, "reward_std": 0.426084041595459, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0, "rewards/format_reward_func/std": 0.0, "step": 361, "step_time": 297.6995941614732 }, { "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": 4096.0, "completions/max_terminated_length": 3144.0, "completions/mean_length": 1132.796875, "completions/mean_terminated_length": 935.2500610351562, "completions/min_length": 171.0, "completions/min_terminated_length": 171.0, "entropy": 3.843835461884737, "epoch": 0.20521541950113378, "frac_reward_zero_std": 0.75, "grad_norm": 0.2099609375, "learning_rate": 7.953514739229025e-07, "loss": -0.0094, "num_tokens": 34361686.0, "reward": 0.109375, "reward_std": 0.37234508991241455, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 362, "step_time": 290.0306559698656 }, { "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": 0.203125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4010.0, "completions/mean_length": 1646.9375, "completions/mean_terminated_length": 1022.6666870117188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.095686696469784, "epoch": 0.20578231292517007, "frac_reward_zero_std": 0.75, "grad_norm": 0.1591796875, "learning_rate": 7.947845804988662e-07, "loss": 0.0258, "num_tokens": 34478306.0, "reward": 0.125, "reward_std": 0.4364357888698578, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 363, "step_time": 292.29391089081764 }, { "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": 4096.0, "completions/max_terminated_length": 3904.0, "completions/mean_length": 794.640625, "completions/mean_terminated_length": 574.550048828125, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 3.039515435695648, "epoch": 0.20634920634920634, "frac_reward_zero_std": 0.625, "grad_norm": 0.34765625, "learning_rate": 7.942176870748299e-07, "loss": -0.004, "num_tokens": 34542979.0, "reward": 0.125, "reward_std": 0.4364357888698578, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 364, "step_time": 301.3254279298708 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3992.0, "completions/mean_length": 1221.8125, "completions/mean_terminated_length": 811.2142944335938, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.787191729992628, "epoch": 0.20691609977324263, "frac_reward_zero_std": 0.625, "grad_norm": 0.26953125, "learning_rate": 7.936507936507936e-07, "loss": 0.0149, "num_tokens": 34631447.0, "reward": 0.1640625, "reward_std": 0.496453195810318, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 365, "step_time": 290.84672499168664 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3070.0, "completions/mean_length": 1016.0, "completions/mean_terminated_length": 697.3793334960938, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.054426599293947, "epoch": 0.20748299319727892, "frac_reward_zero_std": 0.75, "grad_norm": 0.2294921875, "learning_rate": 7.930839002267573e-07, "loss": 0.0136, "num_tokens": 34710095.0, "reward": 0.125, "reward_std": 0.37796446681022644, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 366, "step_time": 304.0007359739393 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3689.0, "completions/mean_length": 1449.328125, "completions/mean_terminated_length": 959.2037353515625, "completions/min_length": 114.0, "completions/min_terminated_length": 114.0, "entropy": 3.523150946944952, "epoch": 0.20804988662131518, "frac_reward_zero_std": 0.875, "grad_norm": 0.142578125, "learning_rate": 7.92517006802721e-07, "loss": 0.0494, "num_tokens": 34814596.0, "reward": 0.0234375, "reward_std": 0.10652101039886475, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 367, "step_time": 293.88083042483777 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3317.0, "completions/mean_length": 1251.921875, "completions/mean_terminated_length": 786.5272216796875, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 3.847768347710371, "epoch": 0.20861678004535147, "frac_reward_zero_std": 0.625, "grad_norm": 0.306640625, "learning_rate": 7.919501133786848e-07, "loss": 0.0848, "num_tokens": 34907127.0, "reward": 0.1015625, "reward_std": 0.3692515790462494, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 368, "step_time": 295.3268093923107 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3967.0, "completions/mean_length": 1433.53125, "completions/mean_terminated_length": 1106.5614013671875, "completions/min_length": 31.0, "completions/min_terminated_length": 31.0, "entropy": 3.2759260311722755, "epoch": 0.20918367346938777, "frac_reward_zero_std": 0.625, "grad_norm": 0.212890625, "learning_rate": 7.913832199546485e-07, "loss": 0.0619, "num_tokens": 35010585.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 369, "step_time": 294.13700341340154 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4038.0, "completions/mean_length": 1461.78125, "completions/mean_terminated_length": 1085.46435546875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.000051070004702, "epoch": 0.20975056689342403, "frac_reward_zero_std": 0.75, "grad_norm": 0.1591796875, "learning_rate": 7.908163265306123e-07, "loss": 0.0527, "num_tokens": 35116339.0, "reward": 0.1796875, "reward_std": 0.5298036932945251, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 370, "step_time": 293.23267390206456 }, { "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": 4096.0, "completions/max_terminated_length": 3713.0, "completions/mean_length": 1095.296875, "completions/mean_terminated_length": 895.2500610351562, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.9960951805114746, "epoch": 0.21031746031746032, "frac_reward_zero_std": 0.875, "grad_norm": 0.1923828125, "learning_rate": 7.902494331065758e-07, "loss": -0.0041, "num_tokens": 35198182.0, "reward": 0.046875, "reward_std": 0.14689241349697113, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 371, "step_time": 292.76586372498423 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3915.0, "completions/mean_length": 1311.28125, "completions/mean_terminated_length": 913.46435546875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.455425687134266, "epoch": 0.2108843537414966, "frac_reward_zero_std": 0.75, "grad_norm": 0.2119140625, "learning_rate": 7.896825396825396e-07, "loss": 0.0223, "num_tokens": 35292584.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 372, "step_time": 290.3787582051009 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3703.0, "completions/mean_length": 956.875, "completions/mean_terminated_length": 690.8474731445312, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 2.503840748220682, "epoch": 0.21145124716553287, "frac_reward_zero_std": 0.75, "grad_norm": 0.318359375, "learning_rate": 7.891156462585033e-07, "loss": -0.002, "num_tokens": 35365000.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 373, "step_time": 290.9691106136888 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3627.0, "completions/mean_length": 1120.859375, "completions/mean_terminated_length": 755.4912109375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.1432966478168964, "epoch": 0.21201814058956917, "frac_reward_zero_std": 0.5, "grad_norm": 0.328125, "learning_rate": 7.885487528344671e-07, "loss": 0.0379, "num_tokens": 35448095.0, "reward": 0.203125, "reward_std": 0.5890451669692993, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 374, "step_time": 294.39544698316604 }, { "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": 4096.0, "completions/max_terminated_length": 3167.0, "completions/mean_length": 1083.828125, "completions/mean_terminated_length": 883.0167236328125, "completions/min_length": 168.0, "completions/min_terminated_length": 168.0, "entropy": 2.5954567454755306, "epoch": 0.21258503401360543, "frac_reward_zero_std": 0.625, "grad_norm": 0.322265625, "learning_rate": 7.879818594104308e-07, "loss": 0.0475, "num_tokens": 35528900.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 375, "step_time": 291.4088851120323 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3934.0, "completions/mean_length": 1236.0, "completions/mean_terminated_length": 884.7719116210938, "completions/min_length": 138.0, "completions/min_terminated_length": 138.0, "entropy": 2.970957178622484, "epoch": 0.21315192743764172, "frac_reward_zero_std": 0.5, "grad_norm": 0.333984375, "learning_rate": 7.874149659863946e-07, "loss": 0.0352, "num_tokens": 35619332.0, "reward": 0.2109375, "reward_std": 0.5896238088607788, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 376, "step_time": 292.5934300655499 }, { "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": 0.015625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3669.0, "completions/mean_length": 739.875, "completions/mean_terminated_length": 686.6032104492188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.692386300303042, "epoch": 0.213718820861678, "frac_reward_zero_std": 0.625, "grad_norm": 0.3515625, "learning_rate": 7.868480725623582e-07, "loss": -0.014, "num_tokens": 35676428.0, "reward": 0.1484375, "reward_std": 0.44257262349128723, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 377, "step_time": 288.1597346281633 }, { "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": 4096.0, "completions/max_terminated_length": 3035.0, "completions/mean_length": 854.25, "completions/mean_terminated_length": 638.1333618164062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.2150022611021996, "epoch": 0.21428571428571427, "frac_reward_zero_std": 0.75, "grad_norm": 0.291015625, "learning_rate": 7.86281179138322e-07, "loss": 0.0244, "num_tokens": 35742540.0, "reward": 0.140625, "reward_std": 0.4917473793029785, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 378, "step_time": 292.05306577309966 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3679.0, "completions/mean_length": 1285.90625, "completions/mean_terminated_length": 826.0726928710938, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 3.13770492374897, "epoch": 0.21485260770975056, "frac_reward_zero_std": 0.75, "grad_norm": 0.2099609375, "learning_rate": 7.857142857142856e-07, "loss": 0.0408, "num_tokens": 35837718.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 379, "step_time": 296.4294180404395 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3570.0, "completions/mean_length": 1455.296875, "completions/mean_terminated_length": 1023.1817626953125, "completions/min_length": 18.0, "completions/min_terminated_length": 18.0, "entropy": 3.678416233509779, "epoch": 0.21541950113378686, "frac_reward_zero_std": 0.625, "grad_norm": 0.2353515625, "learning_rate": 7.851473922902494e-07, "loss": 0.0656, "num_tokens": 35946609.0, "reward": 0.296875, "reward_std": 0.6884463429450989, "rewards/accuracy_reward_func/mean": 0.25, "rewards/accuracy_reward_func/std": 0.6666666865348816, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 380, "step_time": 306.12296494748443 }, { "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": 4096.0, "completions/max_terminated_length": 4057.0, "completions/mean_length": 1195.921875, "completions/mean_terminated_length": 1002.5833740234375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.9083365723490715, "epoch": 0.21598639455782312, "frac_reward_zero_std": 0.625, "grad_norm": 0.291015625, "learning_rate": 7.845804988662131e-07, "loss": 0.1124, "num_tokens": 36034644.0, "reward": 0.2109375, "reward_std": 0.5896238088607788, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 381, "step_time": 304.7264767102897 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3603.0, "completions/mean_length": 965.46875, "completions/mean_terminated_length": 641.6206665039062, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.077329568564892, "epoch": 0.2165532879818594, "frac_reward_zero_std": 0.625, "grad_norm": 0.298828125, "learning_rate": 7.840136054421769e-07, "loss": 0.0352, "num_tokens": 36107762.0, "reward": 0.1328125, "reward_std": 0.4386323094367981, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 382, "step_time": 292.94550040178 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3181.0, "completions/mean_length": 986.28125, "completions/mean_terminated_length": 722.7457885742188, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 3.3939243853092194, "epoch": 0.2171201814058957, "frac_reward_zero_std": 0.375, "grad_norm": 0.470703125, "learning_rate": 7.834467120181405e-07, "loss": 0.1203, "num_tokens": 36183340.0, "reward": 0.1796875, "reward_std": 0.49894481897354126, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 383, "step_time": 304.64304075203836 }, { "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": 4096.0, "completions/max_terminated_length": 2973.0, "completions/mean_length": 926.84375, "completions/mean_terminated_length": 715.5667114257812, "completions/min_length": 1.0, "completions/min_terminated_length": 1.0, "entropy": 2.5942840203642845, "epoch": 0.21768707482993196, "frac_reward_zero_std": 0.75, "grad_norm": 0.3359375, "learning_rate": 7.828798185941043e-07, "loss": 0.0426, "num_tokens": 36256346.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 384, "step_time": 300.3803983433172 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3877.0, "completions/mean_length": 1097.9375, "completions/mean_terminated_length": 787.7930908203125, "completions/min_length": 95.0, "completions/min_terminated_length": 95.0, "entropy": 3.2941271886229515, "epoch": 0.21825396825396826, "frac_reward_zero_std": 0.625, "grad_norm": 0.283203125, "learning_rate": 7.82312925170068e-07, "loss": 0.0227, "num_tokens": 36338230.0, "reward": 0.140625, "reward_std": 0.4917473793029785, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 385, "step_time": 299.5097037758678 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3971.0, "completions/mean_length": 1234.625, "completions/mean_terminated_length": 1093.901611328125, "completions/min_length": 157.0, "completions/min_terminated_length": 157.0, "entropy": 5.1645600609481335, "epoch": 0.21882086167800455, "frac_reward_zero_std": 0.5, "grad_norm": 0.345703125, "learning_rate": 7.817460317460318e-07, "loss": 0.1179, "num_tokens": 36433182.0, "reward": 0.171875, "reward_std": 0.49776285886764526, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 386, "step_time": 319.9222760675475 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 2687.0, "completions/mean_length": 829.875, "completions/mean_terminated_length": 669.245849609375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.8999301306903362, "epoch": 0.2193877551020408, "frac_reward_zero_std": 0.625, "grad_norm": 0.42578125, "learning_rate": 7.811791383219954e-07, "loss": -0.0005, "num_tokens": 36499646.0, "reward": 0.1328125, "reward_std": 0.4386323094367981, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 387, "step_time": 298.64586913306266 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3763.0, "completions/mean_length": 1399.625, "completions/mean_terminated_length": 900.2963256835938, "completions/min_length": 151.0, "completions/min_terminated_length": 151.0, "entropy": 3.2953366115689278, "epoch": 0.2199546485260771, "frac_reward_zero_std": 0.75, "grad_norm": 0.2431640625, "learning_rate": 7.806122448979591e-07, "loss": -0.0318, "num_tokens": 36600422.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 388, "step_time": 292.96577030979097 }, { "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": 0.171875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4075.0, "completions/mean_length": 1561.515625, "completions/mean_terminated_length": 1035.4906005859375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.6118797697126865, "epoch": 0.2205215419501134, "frac_reward_zero_std": 0.75, "grad_norm": 0.234375, "learning_rate": 7.800453514739229e-07, "loss": 0.044, "num_tokens": 36713687.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 389, "step_time": 294.06946489494294 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3862.0, "completions/mean_length": 1409.84375, "completions/mean_terminated_length": 1182.203369140625, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 4.216452792286873, "epoch": 0.22108843537414966, "frac_reward_zero_std": 0.625, "grad_norm": 0.263671875, "learning_rate": 7.794784580498865e-07, "loss": 0.0152, "num_tokens": 36815173.0, "reward": 0.1484375, "reward_std": 0.49344661831855774, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 390, "step_time": 291.6298964070156 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4031.0, "completions/mean_length": 995.765625, "completions/mean_terminated_length": 675.0516967773438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.4762550219893456, "epoch": 0.22165532879818595, "frac_reward_zero_std": 0.625, "grad_norm": 0.3359375, "learning_rate": 7.789115646258503e-07, "loss": 0.1141, "num_tokens": 36890174.0, "reward": 0.09375, "reward_std": 0.36596253514289856, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 391, "step_time": 290.85019321739674 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3624.0, "completions/mean_length": 1162.65625, "completions/mean_terminated_length": 802.4210815429688, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.874783743172884, "epoch": 0.2222222222222222, "frac_reward_zero_std": 0.75, "grad_norm": 0.248046875, "learning_rate": 7.78344671201814e-07, "loss": 0.0289, "num_tokens": 36975984.0, "reward": 0.1875, "reward_std": 0.559903621673584, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 392, "step_time": 290.0920586613938 }, { "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": 0.1875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3583.0, "completions/mean_length": 1432.109375, "completions/mean_terminated_length": 817.3654174804688, "completions/min_length": 96.0, "completions/min_terminated_length": 96.0, "entropy": 2.9061954021453857, "epoch": 0.2227891156462585, "frac_reward_zero_std": 0.875, "grad_norm": 0.10302734375, "learning_rate": 7.777777777777778e-07, "loss": -0.0015, "num_tokens": 37078719.0, "reward": 0.171875, "reward_std": 0.6121699213981628, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 393, "step_time": 292.8030864810571 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3796.0, "completions/mean_length": 1535.53125, "completions/mean_terminated_length": 1061.370361328125, "completions/min_length": 9.0, "completions/min_terminated_length": 9.0, "entropy": 2.842077411711216, "epoch": 0.2233560090702948, "frac_reward_zero_std": 0.75, "grad_norm": 0.1669921875, "learning_rate": 7.772108843537415e-07, "loss": 0.0239, "num_tokens": 37189529.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 394, "step_time": 298.6000028811395 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3664.0, "completions/mean_length": 1395.3125, "completions/mean_terminated_length": 895.1851806640625, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 3.3566896952688694, "epoch": 0.22392290249433106, "frac_reward_zero_std": 0.875, "grad_norm": 0.1142578125, "learning_rate": 7.766439909297052e-07, "loss": -0.0004, "num_tokens": 37290837.0, "reward": 0.046875, "reward_std": 0.14689241349697113, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 395, "step_time": 298.4092497024685 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3760.0, "completions/mean_length": 1193.8125, "completions/mean_terminated_length": 837.4035034179688, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 3.200073454529047, "epoch": 0.22448979591836735, "frac_reward_zero_std": 0.625, "grad_norm": 0.314453125, "learning_rate": 7.760770975056688e-07, "loss": 0.1251, "num_tokens": 37379385.0, "reward": 0.171875, "reward_std": 0.49776285886764526, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 396, "step_time": 301.6528292382136 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4051.0, "completions/mean_length": 1159.890625, "completions/mean_terminated_length": 799.3157958984375, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.7229799032211304, "epoch": 0.22505668934240364, "frac_reward_zero_std": 0.75, "grad_norm": 0.1328125, "learning_rate": 7.755102040816326e-07, "loss": -0.0014, "num_tokens": 37463834.0, "reward": 0.1484375, "reward_std": 0.5092765688896179, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 397, "step_time": 291.51799822505563 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4040.0, "completions/mean_length": 1147.09375, "completions/mean_terminated_length": 897.1864624023438, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.253833882510662, "epoch": 0.2256235827664399, "frac_reward_zero_std": 0.875, "grad_norm": 0.2294921875, "learning_rate": 7.749433106575963e-07, "loss": 0.0462, "num_tokens": 37550696.0, "reward": 0.046875, "reward_std": 0.14689241349697113, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 398, "step_time": 297.5280025154352 }, { "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": 4096.0, "completions/max_terminated_length": 3866.0, "completions/mean_length": 1050.671875, "completions/mean_terminated_length": 847.6500244140625, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 3.5865025259554386, "epoch": 0.2261904761904762, "frac_reward_zero_std": 0.625, "grad_norm": 0.353515625, "learning_rate": 7.743764172335601e-07, "loss": 0.0434, "num_tokens": 37629299.0, "reward": 0.1953125, "reward_std": 0.6275246739387512, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 399, "step_time": 294.2379682837054 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3135.0, "completions/mean_length": 1135.171875, "completions/mean_terminated_length": 712.1964721679688, "completions/min_length": 82.0, "completions/min_terminated_length": 82.0, "entropy": 2.1200684793293476, "epoch": 0.22675736961451248, "frac_reward_zero_std": 0.75, "grad_norm": 0.2421875, "learning_rate": 7.738095238095238e-07, "loss": -0.0039, "num_tokens": 37711574.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 400, "step_time": 286.7584363883361 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 4083.0, "completions/mean_length": 1533.03125, "completions/mean_terminated_length": 1267.8966064453125, "completions/min_length": 166.0, "completions/min_terminated_length": 166.0, "entropy": 5.147508516907692, "epoch": 0.22732426303854875, "frac_reward_zero_std": 0.75, "grad_norm": 0.1923828125, "learning_rate": 7.732426303854876e-07, "loss": 0.0545, "num_tokens": 37822600.0, "reward": 0.078125, "reward_std": 0.2847827076911926, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 401, "step_time": 297.5582052906975 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3628.0, "completions/mean_length": 1282.90625, "completions/mean_terminated_length": 937.4385986328125, "completions/min_length": 44.0, "completions/min_terminated_length": 44.0, "entropy": 3.6668522134423256, "epoch": 0.22789115646258504, "frac_reward_zero_std": 0.875, "grad_norm": 0.1552734375, "learning_rate": 7.726757369614512e-07, "loss": 0.0227, "num_tokens": 37918962.0, "reward": 0.078125, "reward_std": 0.33592739701271057, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 402, "step_time": 299.27116120699793 }, { "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": 0.03125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3322.0, "completions/mean_length": 712.859375, "completions/mean_terminated_length": 603.7257690429688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.6046498268842697, "epoch": 0.22845804988662133, "frac_reward_zero_std": 0.75, "grad_norm": 0.32421875, "learning_rate": 7.721088435374149e-07, "loss": 0.0232, "num_tokens": 37975105.0, "reward": 0.1796875, "reward_std": 0.5445777773857117, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 403, "step_time": 290.56277269963175 }, { "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": 4096.0, "completions/max_terminated_length": 3797.0, "completions/mean_length": 980.75, "completions/mean_terminated_length": 773.0667114257812, "completions/min_length": 46.0, "completions/min_terminated_length": 46.0, "entropy": 2.690210323780775, "epoch": 0.2290249433106576, "frac_reward_zero_std": 0.375, "grad_norm": 0.494140625, "learning_rate": 7.715419501133786e-07, "loss": -0.0237, "num_tokens": 38050729.0, "reward": 0.2578125, "reward_std": 0.654606282711029, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 404, "step_time": 302.6652940651402 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3764.0, "completions/mean_length": 1172.015625, "completions/mean_terminated_length": 693.54541015625, "completions/min_length": 28.0, "completions/min_terminated_length": 28.0, "entropy": 2.875168077647686, "epoch": 0.22959183673469388, "frac_reward_zero_std": 0.625, "grad_norm": 0.2490234375, "learning_rate": 7.709750566893424e-07, "loss": 0.0123, "num_tokens": 38139354.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 405, "step_time": 295.13788509368896 }, { "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": 4096.0, "completions/max_terminated_length": 3766.0, "completions/mean_length": 1332.390625, "completions/mean_terminated_length": 1148.1500244140625, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 4.122431222349405, "epoch": 0.23015873015873015, "frac_reward_zero_std": 0.5, "grad_norm": 0.390625, "learning_rate": 7.704081632653061e-07, "loss": -0.0174, "num_tokens": 38234835.0, "reward": 0.171875, "reward_std": 0.528690755367279, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 406, "step_time": 292.652489118278 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3612.0, "completions/mean_length": 1513.46875, "completions/mean_terminated_length": 1035.2222900390625, "completions/min_length": 192.0, "completions/min_terminated_length": 192.0, "entropy": 3.4942500442266464, "epoch": 0.23072562358276644, "frac_reward_zero_std": 0.75, "grad_norm": 0.212890625, "learning_rate": 7.698412698412699e-07, "loss": 0.0247, "num_tokens": 38344201.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 407, "step_time": 296.21951427217573 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3746.0, "completions/mean_length": 1438.390625, "completions/mean_terminated_length": 946.24072265625, "completions/min_length": 30.0, "completions/min_terminated_length": 30.0, "entropy": 3.4612411949783564, "epoch": 0.23129251700680273, "frac_reward_zero_std": 0.625, "grad_norm": 0.224609375, "learning_rate": 7.692743764172335e-07, "loss": 0.0272, "num_tokens": 38446714.0, "reward": 0.109375, "reward_std": 0.37234508991241455, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 408, "step_time": 290.6038286788389 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 2742.0, "completions/mean_length": 1078.4375, "completions/mean_terminated_length": 822.7118530273438, "completions/min_length": 5.0, "completions/min_terminated_length": 5.0, "entropy": 2.0362658835947514, "epoch": 0.231859410430839, "frac_reward_zero_std": 0.625, "grad_norm": 0.263671875, "learning_rate": 7.687074829931972e-07, "loss": 0.0145, "num_tokens": 38528342.0, "reward": 0.265625, "reward_std": 0.7235755324363708, "rewards/accuracy_reward_func/mean": 0.21875, "rewards/accuracy_reward_func/std": 0.6291528940200806, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 409, "step_time": 294.6687871944159 }, { "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": 4096.0, "completions/max_terminated_length": 3824.0, "completions/mean_length": 1011.203125, "completions/mean_terminated_length": 805.550048828125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.534676004201174, "epoch": 0.23242630385487528, "frac_reward_zero_std": 0.625, "grad_norm": 0.28125, "learning_rate": 7.68140589569161e-07, "loss": 0.0658, "num_tokens": 38607163.0, "reward": 0.1484375, "reward_std": 0.44257262349128723, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 410, "step_time": 316.20966571196914 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3638.0, "completions/mean_length": 1291.890625, "completions/mean_terminated_length": 1001.8103637695312, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "entropy": 3.6028292030096054, "epoch": 0.23299319727891157, "frac_reward_zero_std": 0.625, "grad_norm": 0.26953125, "learning_rate": 7.675736961451246e-07, "loss": 0.0234, "num_tokens": 38701916.0, "reward": 0.1640625, "reward_std": 0.496453195810318, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 411, "step_time": 298.43249408435076 }, { "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": 0.03125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3570.0, "completions/mean_length": 757.171875, "completions/mean_terminated_length": 649.4677124023438, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "entropy": 2.6323740631341934, "epoch": 0.23356009070294784, "frac_reward_zero_std": 0.75, "grad_norm": 0.330078125, "learning_rate": 7.670068027210884e-07, "loss": -0.0178, "num_tokens": 38760799.0, "reward": 0.3359375, "reward_std": 0.7820135354995728, "rewards/accuracy_reward_func/mean": 0.28125, "rewards/accuracy_reward_func/std": 0.7007648944854736, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 412, "step_time": 291.6942321024835 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3752.0, "completions/mean_length": 1136.359375, "completions/mean_terminated_length": 885.5423583984375, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.2796741127967834, "epoch": 0.23412698412698413, "frac_reward_zero_std": 0.75, "grad_norm": 0.291015625, "learning_rate": 7.664399092970521e-07, "loss": 0.059, "num_tokens": 38843726.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 413, "step_time": 288.89864468947053 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3463.0, "completions/mean_length": 1130.796875, "completions/mean_terminated_length": 645.581787109375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 1.6223801504820585, "epoch": 0.23469387755102042, "frac_reward_zero_std": 0.5, "grad_norm": 0.29296875, "learning_rate": 7.658730158730159e-07, "loss": -0.0119, "num_tokens": 38927457.0, "reward": 0.2109375, "reward_std": 0.5896238088607788, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 414, "step_time": 295.61601979192346 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3493.0, "completions/mean_length": 1067.359375, "completions/mean_terminated_length": 754.0516967773438, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 2.890567261725664, "epoch": 0.23526077097505668, "frac_reward_zero_std": 0.75, "grad_norm": 0.2001953125, "learning_rate": 7.653061224489795e-07, "loss": -0.0095, "num_tokens": 39007856.0, "reward": 0.0546875, "reward_std": 0.26899558305740356, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 415, "step_time": 296.26229373179376 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3376.0, "completions/mean_length": 872.984375, "completions/mean_terminated_length": 599.8474731445312, "completions/min_length": 34.0, "completions/min_terminated_length": 34.0, "entropy": 2.055069725960493, "epoch": 0.23582766439909297, "frac_reward_zero_std": 0.75, "grad_norm": 0.24609375, "learning_rate": 7.647392290249433e-07, "loss": -0.0161, "num_tokens": 39077655.0, "reward": 0.3671875, "reward_std": 0.846290647983551, "rewards/accuracy_reward_func/mean": 0.3125, "rewards/accuracy_reward_func/std": 0.7319250702857971, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 416, "step_time": 300.8905012207106 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3284.0, "completions/mean_length": 1009.515625, "completions/mean_terminated_length": 630.4736938476562, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "entropy": 2.7763387747108936, "epoch": 0.23639455782312926, "frac_reward_zero_std": 0.875, "grad_norm": 0.1767578125, "learning_rate": 7.64172335600907e-07, "loss": 0.004, "num_tokens": 39155464.0, "reward": 0.1015625, "reward_std": 0.4470333158969879, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 417, "step_time": 295.4462336972356 }, { "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": 4096.0, "completions/max_terminated_length": 3616.0, "completions/mean_length": 1062.9375, "completions/mean_terminated_length": 860.7333984375, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 4.691736239939928, "epoch": 0.23696145124716553, "frac_reward_zero_std": 0.75, "grad_norm": 0.341796875, "learning_rate": 7.636054421768708e-07, "loss": -0.0148, "num_tokens": 39236092.0, "reward": 0.1015625, "reward_std": 0.3692515790462494, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 418, "step_time": 300.41248774621636 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3550.0, "completions/mean_length": 1022.890625, "completions/mean_terminated_length": 762.4576416015625, "completions/min_length": 64.0, "completions/min_terminated_length": 64.0, "entropy": 3.0089177638292313, "epoch": 0.23752834467120182, "frac_reward_zero_std": 0.75, "grad_norm": 0.2431640625, "learning_rate": 7.630385487528344e-07, "loss": 0.0142, "num_tokens": 39313101.0, "reward": 0.0859375, "reward_std": 0.28942590951919556, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 419, "step_time": 292.61402578186244 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3694.0, "completions/mean_length": 1287.328125, "completions/mean_terminated_length": 886.08935546875, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.8325034826993942, "epoch": 0.23809523809523808, "frac_reward_zero_std": 0.75, "grad_norm": 0.23046875, "learning_rate": 7.624716553287982e-07, "loss": 0.0794, "num_tokens": 39407650.0, "reward": 0.0703125, "reward_std": 0.27984103560447693, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 420, "step_time": 293.98897682875395 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3732.0, "completions/mean_length": 1307.09375, "completions/mean_terminated_length": 850.7272338867188, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 2.6860983073711395, "epoch": 0.23866213151927437, "frac_reward_zero_std": 0.75, "grad_norm": 0.259765625, "learning_rate": 7.619047619047618e-07, "loss": -0.0471, "num_tokens": 39503064.0, "reward": 0.046875, "reward_std": 0.2630521357059479, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 421, "step_time": 292.1864783503115 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3812.0, "completions/mean_length": 1139.359375, "completions/mean_terminated_length": 833.5, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.3448737300932407, "epoch": 0.23922902494331066, "frac_reward_zero_std": 0.625, "grad_norm": 0.27734375, "learning_rate": 7.613378684807256e-07, "loss": 0.0136, "num_tokens": 39586783.0, "reward": 0.15625, "reward_std": 0.5261042714118958, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 422, "step_time": 291.81574575416744 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3708.0, "completions/mean_length": 1263.34375, "completions/mean_terminated_length": 1023.2881469726562, "completions/min_length": 141.0, "completions/min_terminated_length": 141.0, "entropy": 4.240227870643139, "epoch": 0.23979591836734693, "frac_reward_zero_std": 0.625, "grad_norm": 0.306640625, "learning_rate": 7.607709750566893e-07, "loss": -0.0135, "num_tokens": 39679549.0, "reward": 0.078125, "reward_std": 0.3587760329246521, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 423, "step_time": 293.9283110573888 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3438.0, "completions/mean_length": 1057.046875, "completions/mean_terminated_length": 742.6724243164062, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "entropy": 3.5133010260760784, "epoch": 0.24036281179138322, "frac_reward_zero_std": 0.375, "grad_norm": 0.41796875, "learning_rate": 7.602040816326531e-07, "loss": 0.0317, "num_tokens": 39758656.0, "reward": 0.171875, "reward_std": 0.5434950590133667, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 424, "step_time": 292.1548878410831 }, { "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": 0.125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3842.0, "completions/mean_length": 1391.0, "completions/mean_terminated_length": 1004.5714721679688, "completions/min_length": 89.0, "completions/min_terminated_length": 89.0, "entropy": 4.674053099006414, "epoch": 0.2409297052154195, "frac_reward_zero_std": 0.75, "grad_norm": 0.201171875, "learning_rate": 7.596371882086168e-07, "loss": -0.0078, "num_tokens": 39860272.0, "reward": 0.140625, "reward_std": 0.4917473793029785, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.015625, "rewards/format_reward_func/std": 0.08768405020236969, "step": 425, "step_time": 296.3357087178156 }, { "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": 0.140625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3054.0, "completions/mean_length": 1123.359375, "completions/mean_terminated_length": 636.92724609375, "completions/min_length": 120.0, "completions/min_terminated_length": 120.0, "entropy": 2.6381155252456665, "epoch": 0.24149659863945577, "frac_reward_zero_std": 0.75, "grad_norm": 0.203125, "learning_rate": 7.590702947845806e-07, "loss": 0.0053, "num_tokens": 39946383.0, "reward": 0.234375, "reward_std": 0.6170124411582947, "rewards/accuracy_reward_func/mean": 0.1875, "rewards/accuracy_reward_func/std": 0.5875696539878845, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 426, "step_time": 317.2725588269532 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3864.0, "completions/mean_length": 1154.234375, "completions/mean_terminated_length": 849.913818359375, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 2.8968879356980324, "epoch": 0.24206349206349206, "frac_reward_zero_std": 0.75, "grad_norm": 0.19921875, "learning_rate": 7.585034013605441e-07, "loss": 0.0335, "num_tokens": 40033462.0, "reward": 0.2109375, "reward_std": 0.6287092566490173, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0546875, "rewards/format_reward_func/std": 0.15728822350502014, "step": 427, "step_time": 302.59295161347836 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3406.0, "completions/mean_length": 1071.1875, "completions/mean_terminated_length": 814.8474731445312, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.424909345805645, "epoch": 0.24263038548752835, "frac_reward_zero_std": 0.75, "grad_norm": 0.2294921875, "learning_rate": 7.579365079365078e-07, "loss": 0.0391, "num_tokens": 40114154.0, "reward": 0.0625, "reward_std": 0.2745848298072815, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 428, "step_time": 297.0294193122536 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3433.0, "completions/mean_length": 1237.546875, "completions/mean_terminated_length": 941.8448486328125, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.4022247679531574, "epoch": 0.24319727891156462, "frac_reward_zero_std": 0.75, "grad_norm": 0.240234375, "learning_rate": 7.573696145124716e-07, "loss": 0.0661, "num_tokens": 40204189.0, "reward": 0.078125, "reward_std": 0.2847827076911926, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.046875, "rewards/format_reward_func/std": 0.14689241349697113, "step": 429, "step_time": 290.6828445987776 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3396.0, "completions/mean_length": 971.4375, "completions/mean_terminated_length": 706.64404296875, "completions/min_length": 7.0, "completions/min_terminated_length": 7.0, "entropy": 2.267667520791292, "epoch": 0.2437641723356009, "frac_reward_zero_std": 0.875, "grad_norm": 0.171875, "learning_rate": 7.568027210884353e-07, "loss": 0.0079, "num_tokens": 40280201.0, "reward": 0.0703125, "reward_std": 0.33174845576286316, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 430, "step_time": 297.2966445777565 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 3872.0, "completions/mean_length": 966.875, "completions/mean_terminated_length": 812.9835815429688, "completions/min_length": 3.0, "completions/min_terminated_length": 3.0, "entropy": 3.038252253085375, "epoch": 0.2443310657596372, "frac_reward_zero_std": 0.5, "grad_norm": 0.38671875, "learning_rate": 7.562358276643991e-07, "loss": 0.0928, "num_tokens": 40352961.0, "reward": 0.2578125, "reward_std": 0.666620135307312, "rewards/accuracy_reward_func/mean": 0.25, "rewards/accuracy_reward_func/std": 0.6666666865348816, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 431, "step_time": 292.7564950613305 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3765.0, "completions/mean_length": 1053.09375, "completions/mean_terminated_length": 795.2203369140625, "completions/min_length": 8.0, "completions/min_terminated_length": 8.0, "entropy": 2.8713559471070766, "epoch": 0.24489795918367346, "frac_reward_zero_std": 0.375, "grad_norm": 0.404296875, "learning_rate": 7.556689342403628e-07, "loss": 0.0421, "num_tokens": 40430823.0, "reward": 0.3359375, "reward_std": 0.7295067310333252, "rewards/accuracy_reward_func/mean": 0.3125, "rewards/accuracy_reward_func/std": 0.7319250702857971, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 432, "step_time": 291.36802452243865 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3973.0, "completions/mean_length": 1103.515625, "completions/mean_terminated_length": 736.017578125, "completions/min_length": 6.0, "completions/min_terminated_length": 6.0, "entropy": 2.064266886562109, "epoch": 0.24546485260770975, "frac_reward_zero_std": 0.625, "grad_norm": 0.2431640625, "learning_rate": 7.551020408163265e-07, "loss": -0.0609, "num_tokens": 40513864.0, "reward": 0.46875, "reward_std": 0.7961963415145874, "rewards/accuracy_reward_func/mean": 0.40625, "rewards/accuracy_reward_func/std": 0.8110105991363525, "rewards/format_reward_func/mean": 0.0625, "rewards/format_reward_func/std": 0.1666666716337204, "step": 433, "step_time": 308.96737704612315 }, { "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": 0.015625, "completions/max_length": 4096.0, "completions/max_terminated_length": 4004.0, "completions/mean_length": 985.875, "completions/mean_terminated_length": 936.5079956054688, "completions/min_length": 4.0, "completions/min_terminated_length": 4.0, "entropy": 3.8466654792428017, "epoch": 0.24603174603174602, "frac_reward_zero_std": 0.375, "grad_norm": 0.455078125, "learning_rate": 7.545351473922901e-07, "loss": 0.0883, "num_tokens": 40596576.0, "reward": 0.2890625, "reward_std": 0.6655030846595764, "rewards/accuracy_reward_func/mean": 0.25, "rewards/accuracy_reward_func/std": 0.6666666865348816, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 434, "step_time": 343.7383713470772 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3289.0, "completions/mean_length": 1142.4375, "completions/mean_terminated_length": 892.1356201171875, "completions/min_length": 11.0, "completions/min_terminated_length": 11.0, "entropy": 4.053468760102987, "epoch": 0.2465986394557823, "frac_reward_zero_std": 0.625, "grad_norm": 0.33203125, "learning_rate": 7.539682539682539e-07, "loss": 0.0174, "num_tokens": 40680356.0, "reward": 0.15625, "reward_std": 0.49501484632492065, "rewards/accuracy_reward_func/mean": 0.125, "rewards/accuracy_reward_func/std": 0.48795005679130554, "rewards/format_reward_func/mean": 0.03125, "rewards/format_reward_func/std": 0.12198751419782639, "step": 435, "step_time": 296.9095139624551 }, { "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": 0.109375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3774.0, "completions/mean_length": 1235.484375, "completions/mean_terminated_length": 884.1929931640625, "completions/min_length": 13.0, "completions/min_terminated_length": 13.0, "entropy": 3.2999472692608833, "epoch": 0.2471655328798186, "frac_reward_zero_std": 0.75, "grad_norm": 0.2734375, "learning_rate": 7.534013605442176e-07, "loss": 0.0046, "num_tokens": 40769835.0, "reward": 0.0390625, "reward_std": 0.2567298710346222, "rewards/accuracy_reward_func/mean": 0.03125, "rewards/accuracy_reward_func/std": 0.25, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 436, "step_time": 292.59711193386465 }, { "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": 0.03125, "completions/max_length": 4096.0, "completions/max_terminated_length": 3827.0, "completions/mean_length": 953.796875, "completions/mean_terminated_length": 852.4354858398438, "completions/min_length": 12.0, "completions/min_terminated_length": 12.0, "entropy": 2.833001058548689, "epoch": 0.24773242630385486, "frac_reward_zero_std": 0.625, "grad_norm": 0.412109375, "learning_rate": 7.528344671201814e-07, "loss": 0.0739, "num_tokens": 40841910.0, "reward": 0.1015625, "reward_std": 0.3692515790462494, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 437, "step_time": 300.1390373688191 }, { "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": 0.046875, "completions/max_length": 4096.0, "completions/max_terminated_length": 4095.0, "completions/mean_length": 1128.25, "completions/mean_terminated_length": 982.2950439453125, "completions/min_length": 101.0, "completions/min_terminated_length": 101.0, "entropy": 2.0803209394216537, "epoch": 0.24829931972789115, "frac_reward_zero_std": 0.75, "grad_norm": 0.265625, "learning_rate": 7.522675736961451e-07, "loss": 0.027, "num_tokens": 40925342.0, "reward": 0.0859375, "reward_std": 0.36247265338897705, "rewards/accuracy_reward_func/mean": 0.0625, "rewards/accuracy_reward_func/std": 0.35073620080947876, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 438, "step_time": 291.39295391831547 }, { "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": 0.09375, "completions/max_length": 4096.0, "completions/max_terminated_length": 3775.0, "completions/mean_length": 1051.453125, "completions/mean_terminated_length": 736.5, "completions/min_length": 177.0, "completions/min_terminated_length": 177.0, "entropy": 2.9952870681881905, "epoch": 0.24886621315192745, "frac_reward_zero_std": 0.5, "grad_norm": 0.3359375, "learning_rate": 7.517006802721088e-07, "loss": 0.0896, "num_tokens": 41005243.0, "reward": 0.1953125, "reward_std": 0.5463964939117432, "rewards/accuracy_reward_func/mean": 0.15625, "rewards/accuracy_reward_func/std": 0.5409794449806213, "rewards/format_reward_func/mean": 0.0390625, "rewards/format_reward_func/std": 0.13524486124515533, "step": 439, "step_time": 293.7486053816974 }, { "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": 0.078125, "completions/max_length": 4096.0, "completions/max_terminated_length": 4020.0, "completions/mean_length": 1182.953125, "completions/mean_terminated_length": 936.0847778320312, "completions/min_length": 72.0, "completions/min_terminated_length": 72.0, "entropy": 3.639759674668312, "epoch": 0.2494331065759637, "frac_reward_zero_std": 0.625, "grad_norm": 0.271484375, "learning_rate": 7.511337868480725e-07, "loss": 0.0834, "num_tokens": 41092184.0, "reward": 0.1171875, "reward_std": 0.43408530950546265, "rewards/accuracy_reward_func/mean": 0.09375, "rewards/accuracy_reward_func/std": 0.42608407139778137, "rewards/format_reward_func/mean": 0.0234375, "rewards/format_reward_func/std": 0.10652101784944534, "step": 440, "step_time": 297.53792553953826 }, { "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": 0.15625, "completions/max_length": 4096.0, "completions/max_terminated_length": 3768.0, "completions/mean_length": 1477.265625, "completions/mean_terminated_length": 992.3148193359375, "completions/min_length": 2.0, "completions/min_terminated_length": 2.0, "entropy": 3.21871991828084, "epoch": 0.25, "frac_reward_zero_std": 0.875, "grad_norm": 0.1357421875, "learning_rate": 7.505668934240363e-07, "loss": 0.0139, "num_tokens": 41198913.0, "reward": 0.0078125, "reward_std": 0.0625, "rewards/accuracy_reward_func/mean": 0.0, "rewards/accuracy_reward_func/std": 0.0, "rewards/format_reward_func/mean": 0.0078125, "rewards/format_reward_func/std": 0.0625, "step": 441, "step_time": 296.7727650003508 } ], "logging_steps": 1, "max_steps": 1764, "num_input_tokens_seen": 41198913, "num_train_epochs": 1, "save_steps": 500, "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": 4, "trial_name": null, "trial_params": null }