diff --git "a/trainer_state.json" "b/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/trainer_state.json"
@@ -0,0 +1,3943 @@
+{
+  "best_global_step": null,
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 0.0074413989830088055,
+  "eval_steps": 500,
+  "global_step": 300,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7188.5,
+      "epoch": 2.480466327669602e-05,
+      "grad_norm": 0.0,
+      "learning_rate": 3.3333333333333334e-08,
+      "loss": 0.0,
+      "num_tokens": 29342.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 1
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4402.25,
+      "epoch": 4.960932655339204e-05,
+      "grad_norm": 0.0,
+      "learning_rate": 6.666666666666667e-08,
+      "loss": 0.0,
+      "num_tokens": 47507.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 2
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4629.75,
+      "epoch": 7.441398983008806e-05,
+      "grad_norm": 0.0,
+      "learning_rate": 1e-07,
+      "loss": 0.0,
+      "num_tokens": 66606.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 3
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4487.75,
+      "epoch": 9.921865310678408e-05,
+      "grad_norm": 0.0,
+      "learning_rate": 1.3333333333333334e-07,
+      "loss": 0.0,
+      "num_tokens": 85313.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 4
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3864.5,
+      "epoch": 0.00012402331638348009,
+      "grad_norm": 0.0,
+      "learning_rate": 1.6666666666666665e-07,
+      "loss": 0.0,
+      "num_tokens": 101771.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 5
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 756.75,
+      "epoch": 0.00014882797966017612,
+      "grad_norm": 0.0,
+      "learning_rate": 2e-07,
+      "loss": 0.0,
+      "num_tokens": 105186.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 6
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 317.0,
+      "epoch": 0.00017363264293687212,
+      "grad_norm": 0.0,
+      "learning_rate": 2.3333333333333333e-07,
+      "loss": 0.0,
+      "num_tokens": 107054.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 7
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 957.5,
+      "epoch": 0.00019843730621356816,
+      "grad_norm": 0.0,
+      "learning_rate": 2.6666666666666667e-07,
+      "loss": 0.0,
+      "num_tokens": 111668.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 8
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3236.75,
+      "epoch": 0.00022324196949026416,
+      "grad_norm": 0.0,
+      "learning_rate": 3e-07,
+      "loss": 0.0,
+      "num_tokens": 125363.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 9
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4005.0,
+      "epoch": 0.00024804663276696017,
+      "grad_norm": 0.0,
+      "learning_rate": 3.333333333333333e-07,
+      "loss": 0.0,
+      "num_tokens": 141995.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 10
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7189.75,
+      "epoch": 0.00027285129604365623,
+      "grad_norm": 0.0,
+      "learning_rate": 3.666666666666666e-07,
+      "loss": 0.0,
+      "num_tokens": 171334.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 11
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1106.75,
+      "epoch": 0.00029765595932035224,
+      "grad_norm": 0.0,
+      "learning_rate": 4e-07,
+      "loss": 0.0,
+      "num_tokens": 176401.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 12
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7617.0,
+      "epoch": 0.00032246062259704824,
+      "grad_norm": 0.0,
+      "learning_rate": 4.3333333333333335e-07,
+      "loss": 0.0,
+      "num_tokens": 207361.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 13
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 674.25,
+      "epoch": 0.00034726528587374425,
+      "grad_norm": 0.0,
+      "learning_rate": 4.6666666666666666e-07,
+      "loss": 0.0,
+      "num_tokens": 210422.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 14
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7167.25,
+      "epoch": 0.0003720699491504403,
+      "grad_norm": 0.0,
+      "learning_rate": 5e-07,
+      "loss": 0.0,
+      "num_tokens": 239887.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 15
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 756.25,
+      "epoch": 0.0003968746124271363,
+      "grad_norm": 0.0,
+      "learning_rate": 5.333333333333333e-07,
+      "loss": 0.0,
+      "num_tokens": 243312.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 16
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4135.25,
+      "epoch": 0.0004216792757038323,
+      "grad_norm": 0.0,
+      "learning_rate": 5.666666666666666e-07,
+      "loss": 0.0,
+      "num_tokens": 260569.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 17
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4373.0,
+      "epoch": 0.00044648393898052833,
+      "grad_norm": 0.0,
+      "learning_rate": 6e-07,
+      "loss": 0.0,
+      "num_tokens": 278469.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 18
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1157.5,
+      "epoch": 0.00047128860225722433,
+      "grad_norm": 0.0,
+      "learning_rate": 6.333333333333332e-07,
+      "loss": 0.0,
+      "num_tokens": 283755.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 19
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 951.25,
+      "epoch": 0.0004960932655339203,
+      "grad_norm": 0.0,
+      "learning_rate": 6.666666666666666e-07,
+      "loss": 0.0,
+      "num_tokens": 288076.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 20
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 724.5,
+      "epoch": 0.0005208979288106164,
+      "grad_norm": 0.0,
+      "learning_rate": 7e-07,
+      "loss": 0.0,
+      "num_tokens": 291390.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 21
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 657.25,
+      "epoch": 0.0005457025920873125,
+      "grad_norm": 0.0,
+      "learning_rate": 7.333333333333332e-07,
+      "loss": 0.0,
+      "num_tokens": 294403.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 22
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 669.5,
+      "epoch": 0.0005705072553640084,
+      "grad_norm": 0.0,
+      "learning_rate": 7.666666666666667e-07,
+      "loss": 0.0,
+      "num_tokens": 297541.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 23
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3391.0,
+      "epoch": 0.0005953119186407045,
+      "grad_norm": 0.0,
+      "learning_rate": 8e-07,
+      "loss": 0.0,
+      "num_tokens": 311721.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 24
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 881.75,
+      "epoch": 0.0006201165819174004,
+      "grad_norm": 0.0,
+      "learning_rate": 8.333333333333333e-07,
+      "loss": 0.0,
+      "num_tokens": 315736.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 25
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 651.75,
+      "epoch": 0.0006449212451940965,
+      "grad_norm": 0.0,
+      "learning_rate": 8.666666666666667e-07,
+      "loss": 0.0,
+      "num_tokens": 318767.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 26
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1553.0,
+      "epoch": 0.0006697259084707925,
+      "grad_norm": 0.0,
+      "learning_rate": 9e-07,
+      "loss": 0.0,
+      "num_tokens": 325395.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 27
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4422.25,
+      "epoch": 0.0006945305717474885,
+      "grad_norm": 0.0,
+      "learning_rate": 9.333333333333333e-07,
+      "loss": 0.0,
+      "num_tokens": 343972.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 28
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 834.75,
+      "epoch": 0.0007193352350241846,
+      "grad_norm": 0.0,
+      "learning_rate": 9.666666666666666e-07,
+      "loss": 0.0,
+      "num_tokens": 347987.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 29
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4034.25,
+      "epoch": 0.0007441398983008806,
+      "grad_norm": 0.0,
+      "learning_rate": 1e-06,
+      "loss": 0.0,
+      "num_tokens": 364944.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 30
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 789.25,
+      "epoch": 0.0007689445615775766,
+      "grad_norm": 0.0,
+      "learning_rate": 9.99966154001881e-07,
+      "loss": 0.0,
+      "num_tokens": 368653.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 31
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6620.5,
+      "epoch": 0.0007937492248542726,
+      "grad_norm": 0.16441772878170013,
+      "learning_rate": 9.998646205897307e-07,
+      "loss": 0.178,
+      "num_tokens": 395735.0,
+      "reward": 0.25,
+      "reward_std": 0.5,
+      "rewards/accuracy_reward": 0.25,
+      "step": 32
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4088.25,
+      "epoch": 0.0008185538881309686,
+      "grad_norm": 0.0,
+      "learning_rate": 9.996954135095478e-07,
+      "loss": 0.0,
+      "num_tokens": 413020.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 33
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6849.0,
+      "epoch": 0.0008433585514076646,
+      "grad_norm": 0.0,
+      "learning_rate": 9.994585556692624e-07,
+      "loss": 0.0,
+      "num_tokens": 440984.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 34
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1422.75,
+      "epoch": 0.0008681632146843607,
+      "grad_norm": 0.0,
+      "learning_rate": 9.991540791356342e-07,
+      "loss": 0.0,
+      "num_tokens": 447163.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 35
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 556.25,
+      "epoch": 0.0008929678779610567,
+      "grad_norm": 0.0,
+      "learning_rate": 9.98782025129912e-07,
+      "loss": 0.0,
+      "num_tokens": 450580.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 36
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 721.75,
+      "epoch": 0.0009177725412377527,
+      "grad_norm": 0.0,
+      "learning_rate": 9.983424440222529e-07,
+      "loss": 0.0,
+      "num_tokens": 453967.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 37
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2892.5,
+      "epoch": 0.0009425772045144487,
+      "grad_norm": 0.0,
+      "learning_rate": 9.978353953249021e-07,
+      "loss": 0.0,
+      "num_tokens": 466049.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 38
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2498.25,
+      "epoch": 0.0009673818677911447,
+      "grad_norm": 0.0,
+      "learning_rate": 9.972609476841365e-07,
+      "loss": 0.0,
+      "num_tokens": 476494.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 39
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5306.25,
+      "epoch": 0.0009921865310678407,
+      "grad_norm": 0.0,
+      "learning_rate": 9.966191788709714e-07,
+      "loss": 0.0,
+      "num_tokens": 498639.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 40
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2374.25,
+      "epoch": 0.0010169911943445368,
+      "grad_norm": 0.0,
+      "learning_rate": 9.959101757706308e-07,
+      "loss": 0.0,
+      "num_tokens": 508772.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 41
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1571.75,
+      "epoch": 0.0010417958576212328,
+      "grad_norm": 0.0,
+      "learning_rate": 9.95134034370785e-07,
+      "loss": 0.0,
+      "num_tokens": 515583.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 42
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3794.5,
+      "epoch": 0.0010666005208979288,
+      "grad_norm": 0.0,
+      "learning_rate": 9.942908597485558e-07,
+      "loss": 0.0,
+      "num_tokens": 531425.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 43
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2603.25,
+      "epoch": 0.001091405184174625,
+      "grad_norm": 0.0,
+      "learning_rate": 9.933807660562896e-07,
+      "loss": 0.0,
+      "num_tokens": 542478.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 44
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 825.5,
+      "epoch": 0.0011162098474513209,
+      "grad_norm": 0.0,
+      "learning_rate": 9.92403876506104e-07,
+      "loss": 0.0,
+      "num_tokens": 546324.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 45
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 8192.0,
+      "epoch": 0.0011410145107280168,
+      "grad_norm": 0.0,
+      "learning_rate": 9.913603233532067e-07,
+      "loss": 0.0,
+      "num_tokens": 579840.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 46
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3327.0,
+      "epoch": 0.0011658191740047128,
+      "grad_norm": 0.0,
+      "learning_rate": 9.902502478779896e-07,
+      "loss": 0.0,
+      "num_tokens": 593964.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 47
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2645.75,
+      "epoch": 0.001190623837281409,
+      "grad_norm": 0.0,
+      "learning_rate": 9.890738003669027e-07,
+      "loss": 0.0,
+      "num_tokens": 605147.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 48
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2478.75,
+      "epoch": 0.001215428500558105,
+      "grad_norm": 0.0,
+      "learning_rate": 9.878311400921072e-07,
+      "loss": 0.0,
+      "num_tokens": 615638.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 49
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6159.0,
+      "epoch": 0.0012402331638348009,
+      "grad_norm": 0.0,
+      "learning_rate": 9.865224352899118e-07,
+      "loss": 0.0,
+      "num_tokens": 640798.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 50
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3449.25,
+      "epoch": 0.001265037827111497,
+      "grad_norm": 0.0,
+      "learning_rate": 9.851478631379982e-07,
+      "loss": 0.0,
+      "num_tokens": 655443.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 51
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7229.75,
+      "epoch": 0.001289842490388193,
+      "grad_norm": 0.0,
+      "learning_rate": 9.837076097314318e-07,
+      "loss": 0.0,
+      "num_tokens": 685146.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 52
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3381.75,
+      "epoch": 0.001314647153664889,
+      "grad_norm": 0.0,
+      "learning_rate": 9.822018700574694e-07,
+      "loss": 0.0,
+      "num_tokens": 699265.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 53
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6892.0,
+      "epoch": 0.001339451816941585,
+      "grad_norm": 0.0,
+      "learning_rate": 9.806308479691594e-07,
+      "loss": 0.0,
+      "num_tokens": 727501.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 54
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5287.75,
+      "epoch": 0.001364256480218281,
+      "grad_norm": 0.0,
+      "learning_rate": 9.789947561577443e-07,
+      "loss": 0.0,
+      "num_tokens": 749364.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 55
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5510.75,
+      "epoch": 0.001389061143494977,
+      "grad_norm": 0.0,
+      "learning_rate": 9.77293816123866e-07,
+      "loss": 0.0,
+      "num_tokens": 771947.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 56
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1866.25,
+      "epoch": 0.0014138658067716732,
+      "grad_norm": 0.0,
+      "learning_rate": 9.755282581475767e-07,
+      "loss": 0.0,
+      "num_tokens": 780588.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 57
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7529.75,
+      "epoch": 0.0014386704700483691,
+      "grad_norm": 0.0,
+      "learning_rate": 9.736983212571645e-07,
+      "loss": 0.0,
+      "num_tokens": 811727.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 58
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 8192.0,
+      "epoch": 0.001463475133325065,
+      "grad_norm": 0.0,
+      "learning_rate": 9.718042531967916e-07,
+      "loss": 0.0,
+      "num_tokens": 844919.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 59
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4105.0,
+      "epoch": 0.0014882797966017612,
+      "grad_norm": 0.0,
+      "learning_rate": 9.698463103929541e-07,
+      "loss": 0.0,
+      "num_tokens": 861815.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 60
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2881.25,
+      "epoch": 0.0015130844598784572,
+      "grad_norm": 0.0,
+      "learning_rate": 9.678247579197658e-07,
+      "loss": 0.0,
+      "num_tokens": 873944.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 61
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2342.5,
+      "epoch": 0.0015378891231551531,
+      "grad_norm": 0.0,
+      "learning_rate": 9.657398694630712e-07,
+      "loss": 0.0,
+      "num_tokens": 884086.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 62
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7138.75,
+      "epoch": 0.001562693786431849,
+      "grad_norm": 0.0,
+      "learning_rate": 9.635919272833937e-07,
+      "loss": 0.0,
+      "num_tokens": 913165.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 63
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7349.25,
+      "epoch": 0.0015874984497085453,
+      "grad_norm": 0.0,
+      "learning_rate": 9.613812221777212e-07,
+      "loss": 0.0,
+      "num_tokens": 943130.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 64
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3842.5,
+      "epoch": 0.0016123031129852412,
+      "grad_norm": 0.0,
+      "learning_rate": 9.591080534401371e-07,
+      "loss": 0.0,
+      "num_tokens": 959160.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 65
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1762.5,
+      "epoch": 0.0016371077762619372,
+      "grad_norm": 0.0,
+      "learning_rate": 9.567727288213004e-07,
+      "loss": 0.0,
+      "num_tokens": 966626.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 66
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5473.25,
+      "epoch": 0.0016619124395386333,
+      "grad_norm": 0.0,
+      "learning_rate": 9.543755644867822e-07,
+      "loss": 0.0,
+      "num_tokens": 989395.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 67
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5421.75,
+      "epoch": 0.0016867171028153293,
+      "grad_norm": 0.0,
+      "learning_rate": 9.519168849742602e-07,
+      "loss": 0.0,
+      "num_tokens": 1011750.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 68
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5362.0,
+      "epoch": 0.0017115217660920252,
+      "grad_norm": 0.0,
+      "learning_rate": 9.493970231495834e-07,
+      "loss": 0.0,
+      "num_tokens": 1033958.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 69
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1578.0,
+      "epoch": 0.0017363264293687214,
+      "grad_norm": 0.0,
+      "learning_rate": 9.468163201617061e-07,
+      "loss": 0.0,
+      "num_tokens": 1040734.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 70
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3258.5,
+      "epoch": 0.0017611310926454174,
+      "grad_norm": 0.0,
+      "learning_rate": 9.44175125396502e-07,
+      "loss": 0.0,
+      "num_tokens": 1054364.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 71
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1447.0,
+      "epoch": 0.0017859357559221133,
+      "grad_norm": 0.0,
+      "learning_rate": 9.414737964294634e-07,
+      "loss": 0.0,
+      "num_tokens": 1061444.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 72
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4742.25,
+      "epoch": 0.0018107404191988095,
+      "grad_norm": 0.26786288619041443,
+      "learning_rate": 9.387126989772909e-07,
+      "loss": -0.0049,
+      "num_tokens": 1080881.0,
+      "reward": 0.5,
+      "reward_std": 0.5773502588272095,
+      "rewards/accuracy_reward": 0.5,
+      "step": 73
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3059.0,
+      "epoch": 0.0018355450824755054,
+      "grad_norm": 0.0,
+      "learning_rate": 9.358922068483811e-07,
+      "loss": 0.0,
+      "num_tokens": 1093829.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 74
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2770.5,
+      "epoch": 0.0018603497457522014,
+      "grad_norm": 0.0,
+      "learning_rate": 9.330127018922193e-07,
+      "loss": 0.0,
+      "num_tokens": 1105535.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 75
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 8192.0,
+      "epoch": 0.0018851544090288973,
+      "grad_norm": 0.0,
+      "learning_rate": 9.300745739476828e-07,
+      "loss": 0.0,
+      "num_tokens": 1139031.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 76
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6549.5,
+      "epoch": 0.0019099590723055935,
+      "grad_norm": 0.0,
+      "learning_rate": 9.270782207902627e-07,
+      "loss": 0.0,
+      "num_tokens": 1166325.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 77
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 903.25,
+      "epoch": 0.0019347637355822895,
+      "grad_norm": 0.0,
+      "learning_rate": 9.240240480782129e-07,
+      "loss": 0.0,
+      "num_tokens": 1170538.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 78
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2268.0,
+      "epoch": 0.0019595683988589854,
+      "grad_norm": 0.0,
+      "learning_rate": 9.209124692976287e-07,
+      "loss": 0.0,
+      "num_tokens": 1180026.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 79
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4045.25,
+      "epoch": 0.0019843730621356814,
+      "grad_norm": 0.0,
+      "learning_rate": 9.177439057064682e-07,
+      "loss": 0.0,
+      "num_tokens": 1196683.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 80
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4910.75,
+      "epoch": 0.0020091777254123773,
+      "grad_norm": 0.0,
+      "learning_rate": 9.145187862775208e-07,
+      "loss": 0.0,
+      "num_tokens": 1216882.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 81
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 441.75,
+      "epoch": 0.0020339823886890737,
+      "grad_norm": 0.0,
+      "learning_rate": 9.112375476403311e-07,
+      "loss": 0.0,
+      "num_tokens": 1219041.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 82
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6115.75,
+      "epoch": 0.0020587870519657697,
+      "grad_norm": 0.0,
+      "learning_rate": 9.079006340220861e-07,
+      "loss": 0.0,
+      "num_tokens": 1244036.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 83
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 851.0,
+      "epoch": 0.0020835917152424656,
+      "grad_norm": 0.0,
+      "learning_rate": 9.045084971874737e-07,
+      "loss": 0.0,
+      "num_tokens": 1248104.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 84
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1183.5,
+      "epoch": 0.0021083963785191616,
+      "grad_norm": 0.0,
+      "learning_rate": 9.010615963775219e-07,
+      "loss": 0.0,
+      "num_tokens": 1253298.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 85
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2500.75,
+      "epoch": 0.0021332010417958575,
+      "grad_norm": 0.0,
+      "learning_rate": 8.975603982474238e-07,
+      "loss": 0.0,
+      "num_tokens": 1263869.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 86
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6133.0,
+      "epoch": 0.0021580057050725535,
+      "grad_norm": 0.0,
+      "learning_rate": 8.940053768033608e-07,
+      "loss": 0.0,
+      "num_tokens": 1289193.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 87
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1037.25,
+      "epoch": 0.00218281036834925,
+      "grad_norm": 0.0,
+      "learning_rate": 8.903970133383296e-07,
+      "loss": 0.0,
+      "num_tokens": 1293746.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 88
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7457.75,
+      "epoch": 0.002207615031625946,
+      "grad_norm": 0.0,
+      "learning_rate": 8.86735796366982e-07,
+      "loss": 0.0,
+      "num_tokens": 1324261.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 89
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 614.75,
+      "epoch": 0.0022324196949026417,
+      "grad_norm": 0.0,
+      "learning_rate": 8.83022221559489e-07,
+      "loss": 0.0,
+      "num_tokens": 1327344.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 90
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1136.25,
+      "epoch": 0.0022572243581793377,
+      "grad_norm": 0.0,
+      "learning_rate": 8.792567916744345e-07,
+      "loss": 0.0,
+      "num_tokens": 1332425.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 91
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3865.75,
+      "epoch": 0.0022820290214560337,
+      "grad_norm": 0.0,
+      "learning_rate": 8.754400164907496e-07,
+      "loss": 0.0,
+      "num_tokens": 1348472.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 92
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7174.25,
+      "epoch": 0.0023068336847327296,
+      "grad_norm": 0.0,
+      "learning_rate": 8.71572412738697e-07,
+      "loss": 0.0,
+      "num_tokens": 1377825.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 93
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4552.75,
+      "epoch": 0.0023316383480094256,
+      "grad_norm": 0.0,
+      "learning_rate": 8.676545040299143e-07,
+      "loss": 0.0,
+      "num_tokens": 1396588.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 94
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4767.25,
+      "epoch": 0.002356443011286122,
+      "grad_norm": 0.0,
+      "learning_rate": 8.636868207865243e-07,
+      "loss": 0.0,
+      "num_tokens": 1416321.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 95
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1088.75,
+      "epoch": 0.002381247674562818,
+      "grad_norm": 0.0,
+      "learning_rate": 8.596699001693255e-07,
+      "loss": 0.0,
+      "num_tokens": 1421132.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 96
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 8192.0,
+      "epoch": 0.002406052337839514,
+      "grad_norm": 0.0,
+      "learning_rate": 8.556042860050685e-07,
+      "loss": 0.0,
+      "num_tokens": 1454580.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 97
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3769.75,
+      "epoch": 0.00243085700111621,
+      "grad_norm": 0.0,
+      "learning_rate": 8.514905287128309e-07,
+      "loss": 0.0,
+      "num_tokens": 1470555.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 98
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 560.75,
+      "epoch": 0.0024556616643929058,
+      "grad_norm": 0.0,
+      "learning_rate": 8.473291852294986e-07,
+      "loss": 0.0,
+      "num_tokens": 1473302.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 99
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7399.75,
+      "epoch": 0.0024804663276696017,
+      "grad_norm": 0.0,
+      "learning_rate": 8.431208189343669e-07,
+      "loss": 0.0,
+      "num_tokens": 1503753.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 100
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5177.75,
+      "epoch": 0.002505270990946298,
+      "grad_norm": 0.0,
+      "learning_rate": 8.388659995728662e-07,
+      "loss": 0.0,
+      "num_tokens": 1524904.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 101
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5838.5,
+      "epoch": 0.002530075654222994,
+      "grad_norm": 0.0,
+      "learning_rate": 8.34565303179429e-07,
+      "loss": 0.0,
+      "num_tokens": 1548710.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 102
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3594.5,
+      "epoch": 0.00255488031749969,
+      "grad_norm": 0.0,
+      "learning_rate": 8.302193119995038e-07,
+      "loss": 0.0,
+      "num_tokens": 1563572.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 103
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6636.5,
+      "epoch": 0.002579684980776386,
+      "grad_norm": 0.0,
+      "learning_rate": 8.258286144107276e-07,
+      "loss": 0.0,
+      "num_tokens": 1590698.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 104
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3181.25,
+      "epoch": 0.002604489644053082,
+      "grad_norm": 0.0,
+      "learning_rate": 8.213938048432696e-07,
+      "loss": 0.0,
+      "num_tokens": 1603955.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 105
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4676.0,
+      "epoch": 0.002629294307329778,
+      "grad_norm": 0.0,
+      "learning_rate": 8.16915483699355e-07,
+      "loss": 0.0,
+      "num_tokens": 1623231.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 106
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 694.75,
+      "epoch": 0.0026540989706064742,
+      "grad_norm": 0.0,
+      "learning_rate": 8.123942572719799e-07,
+      "loss": 0.0,
+      "num_tokens": 1626562.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 107
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1234.75,
+      "epoch": 0.00267890363388317,
+      "grad_norm": 0.0,
+      "learning_rate": 8.07830737662829e-07,
+      "loss": 0.0,
+      "num_tokens": 1631929.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 108
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2437.0,
+      "epoch": 0.002703708297159866,
+      "grad_norm": 0.0,
+      "learning_rate": 8.032255426994068e-07,
+      "loss": 0.0,
+      "num_tokens": 1642117.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 109
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4050.75,
+      "epoch": 0.002728512960436562,
+      "grad_norm": 0.2529246211051941,
+      "learning_rate": 7.985792958513931e-07,
+      "loss": -0.0316,
+      "num_tokens": 1658812.0,
+      "reward": 0.25,
+      "reward_std": 0.5,
+      "rewards/accuracy_reward": 0.25,
+      "step": 110
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2922.25,
+      "epoch": 0.002753317623713258,
+      "grad_norm": 0.0,
+      "learning_rate": 7.938926261462365e-07,
+      "loss": 0.0,
+      "num_tokens": 1671017.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 111
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5514.75,
+      "epoch": 0.002778122286989954,
+      "grad_norm": 0.2133975625038147,
+      "learning_rate": 7.891661680839932e-07,
+      "loss": 0.1083,
+      "num_tokens": 1694024.0,
+      "reward": 0.25,
+      "reward_std": 0.5,
+      "rewards/accuracy_reward": 0.25,
+      "step": 112
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1548.25,
+      "epoch": 0.00280292695026665,
+      "grad_norm": 0.0,
+      "learning_rate": 7.844005615514258e-07,
+      "loss": 0.0,
+      "num_tokens": 1700829.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 113
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4555.25,
+      "epoch": 0.0028277316135433463,
+      "grad_norm": 0.0,
+      "learning_rate": 7.795964517353733e-07,
+      "loss": 0.0,
+      "num_tokens": 1719742.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 114
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 549.25,
+      "epoch": 0.0028525362768200423,
+      "grad_norm": 0.0,
+      "learning_rate": 7.74754489035403e-07,
+      "loss": 0.0,
+      "num_tokens": 1722415.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 115
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6851.0,
+      "epoch": 0.0028773409400967382,
+      "grad_norm": 0.0,
+      "learning_rate": 7.698753289757564e-07,
+      "loss": 0.0,
+      "num_tokens": 1750299.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 116
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6247.5,
+      "epoch": 0.002902145603373434,
+      "grad_norm": 0.0,
+      "learning_rate": 7.649596321166024e-07,
+      "loss": 0.0,
+      "num_tokens": 1776377.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 117
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6994.5,
+      "epoch": 0.00292695026665013,
+      "grad_norm": 0.0,
+      "learning_rate": 7.600080639646076e-07,
+      "loss": 0.0,
+      "num_tokens": 1805067.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 118
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5748.25,
+      "epoch": 0.002951754929926826,
+      "grad_norm": 0.0,
+      "learning_rate": 7.550212948828376e-07,
+      "loss": 0.0,
+      "num_tokens": 1828768.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 119
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 704.25,
+      "epoch": 0.0029765595932035225,
+      "grad_norm": 0.0,
+      "learning_rate": 7.5e-07,
+      "loss": 0.0,
+      "num_tokens": 1831957.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 120
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2971.25,
+      "epoch": 0.0030013642564802184,
+      "grad_norm": 0.0,
+      "learning_rate": 7.449448591190434e-07,
+      "loss": 0.0,
+      "num_tokens": 1844478.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 121
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1162.5,
+      "epoch": 0.0030261689197569144,
+      "grad_norm": 0.0,
+      "learning_rate": 7.398565566251232e-07,
+      "loss": 0.0,
+      "num_tokens": 1850380.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 122
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 653.75,
+      "epoch": 0.0030509735830336103,
+      "grad_norm": 0.0,
+      "learning_rate": 7.347357813929454e-07,
+      "loss": 0.0,
+      "num_tokens": 1853379.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 123
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 8192.0,
+      "epoch": 0.0030757782463103063,
+      "grad_norm": 0.0,
+      "learning_rate": 7.295832266935058e-07,
+      "loss": 0.0,
+      "num_tokens": 1887027.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 124
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 867.25,
+      "epoch": 0.0031005829095870022,
+      "grad_norm": 0.0,
+      "learning_rate": 7.243995901002311e-07,
+      "loss": 0.0,
+      "num_tokens": 1890888.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 125
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2747.25,
+      "epoch": 0.003125387572863698,
+      "grad_norm": 0.0,
+      "learning_rate": 7.191855733945386e-07,
+      "loss": 0.0,
+      "num_tokens": 1902673.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 126
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 869.75,
+      "epoch": 0.0031501922361403946,
+      "grad_norm": 0.0,
+      "learning_rate": 7.139418824708271e-07,
+      "loss": 0.0,
+      "num_tokens": 1906624.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 127
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 826.5,
+      "epoch": 0.0031749968994170905,
+      "grad_norm": 0.0,
+      "learning_rate": 7.086692272409089e-07,
+      "loss": 0.0,
+      "num_tokens": 1910582.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 128
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5324.75,
+      "epoch": 0.0031998015626937865,
+      "grad_norm": 0.0,
+      "learning_rate": 7.033683215379002e-07,
+      "loss": 0.0,
+      "num_tokens": 1932469.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 129
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1146.75,
+      "epoch": 0.0032246062259704824,
+      "grad_norm": 0.0,
+      "learning_rate": 6.980398830195784e-07,
+      "loss": 0.0,
+      "num_tokens": 1937920.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 130
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 595.0,
+      "epoch": 0.0032494108892471784,
+      "grad_norm": 0.0,
+      "learning_rate": 6.926846330712241e-07,
+      "loss": 0.0,
+      "num_tokens": 1940816.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 131
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2312.5,
+      "epoch": 0.0032742155525238743,
+      "grad_norm": 0.0,
+      "learning_rate": 6.87303296707956e-07,
+      "loss": 0.0,
+      "num_tokens": 1951390.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 132
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1446.5,
+      "epoch": 0.0032990202158005707,
+      "grad_norm": 0.0,
+      "learning_rate": 6.818966024765757e-07,
+      "loss": 0.0,
+      "num_tokens": 1957664.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 133
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1437.0,
+      "epoch": 0.0033238248790772667,
+      "grad_norm": 0.0,
+      "learning_rate": 6.764652823569343e-07,
+      "loss": 0.0,
+      "num_tokens": 1963868.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 134
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3721.25,
+      "epoch": 0.0033486295423539626,
+      "grad_norm": 0.0,
+      "learning_rate": 6.710100716628344e-07,
+      "loss": 0.0,
+      "num_tokens": 1979441.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 135
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2822.25,
+      "epoch": 0.0033734342056306586,
+      "grad_norm": 0.0,
+      "learning_rate": 6.65531708942479e-07,
+      "loss": 0.0,
+      "num_tokens": 1991198.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 136
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7604.75,
+      "epoch": 0.0033982388689073545,
+      "grad_norm": 0.0,
+      "learning_rate": 6.600309358784857e-07,
+      "loss": 0.0,
+      "num_tokens": 2022213.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 137
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1285.75,
+      "epoch": 0.0034230435321840505,
+      "grad_norm": 0.0,
+      "learning_rate": 6.545084971874736e-07,
+      "loss": 0.0,
+      "num_tokens": 2027864.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 138
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2961.0,
+      "epoch": 0.0034478481954607464,
+      "grad_norm": 0.0,
+      "learning_rate": 6.489651405192409e-07,
+      "loss": 0.0,
+      "num_tokens": 2040244.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 139
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3632.5,
+      "epoch": 0.003472652858737443,
+      "grad_norm": 0.0,
+      "learning_rate": 6.434016163555451e-07,
+      "loss": 0.0,
+      "num_tokens": 2055550.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 140
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1019.0,
+      "epoch": 0.0034974575220141388,
+      "grad_norm": 0.0,
+      "learning_rate": 6.378186779084995e-07,
+      "loss": 0.0,
+      "num_tokens": 2060194.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 141
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 8192.0,
+      "epoch": 0.0035222621852908347,
+      "grad_norm": 0.0,
+      "learning_rate": 6.322170810186011e-07,
+      "loss": 0.0,
+      "num_tokens": 2093590.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 142
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 680.75,
+      "epoch": 0.0035470668485675307,
+      "grad_norm": 0.0,
+      "learning_rate": 6.265975840524009e-07,
+      "loss": 0.0,
+      "num_tokens": 2096849.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 143
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3868.5,
+      "epoch": 0.0035718715118442266,
+      "grad_norm": 0.3706275522708893,
+      "learning_rate": 6.209609477998338e-07,
+      "loss": 0.1572,
+      "num_tokens": 2112955.0,
+      "reward": 0.5,
+      "reward_std": 0.5773502588272095,
+      "rewards/accuracy_reward": 0.5,
+      "step": 144
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2470.5,
+      "epoch": 0.0035966761751209226,
+      "grad_norm": 0.0,
+      "learning_rate": 6.153079353712201e-07,
+      "loss": 0.0,
+      "num_tokens": 2123349.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 145
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1336.25,
+      "epoch": 0.003621480838397619,
+      "grad_norm": 0.0,
+      "learning_rate": 6.096393120939516e-07,
+      "loss": 0.0,
+      "num_tokens": 2129378.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 146
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3586.0,
+      "epoch": 0.003646285501674315,
+      "grad_norm": 0.2838553190231323,
+      "learning_rate": 6.039558454088795e-07,
+      "loss": -0.2103,
+      "num_tokens": 2144430.0,
+      "reward": 0.5,
+      "reward_std": 0.5773502588272095,
+      "rewards/accuracy_reward": 0.5,
+      "step": 147
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1188.0,
+      "epoch": 0.003671090164951011,
+      "grad_norm": 0.0,
+      "learning_rate": 5.98258304766415e-07,
+      "loss": 0.0,
+      "num_tokens": 2149782.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 148
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 516.0,
+      "epoch": 0.003695894828227707,
+      "grad_norm": 0.0,
+      "learning_rate": 5.925474615223572e-07,
+      "loss": 0.0,
+      "num_tokens": 2152310.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 149
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2670.25,
+      "epoch": 0.0037206994915044028,
+      "grad_norm": 0.2669675350189209,
+      "learning_rate": 5.868240888334652e-07,
+      "loss": -0.0001,
+      "num_tokens": 2163647.0,
+      "reward": 0.75,
+      "reward_std": 0.5,
+      "rewards/accuracy_reward": 0.75,
+      "step": 150
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 8192.0,
+      "epoch": 0.0037455041547810987,
+      "grad_norm": 0.0,
+      "learning_rate": 5.810889615527838e-07,
+      "loss": 0.0,
+      "num_tokens": 2196923.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 151
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1642.25,
+      "epoch": 0.0037703088180577947,
+      "grad_norm": 0.3528982996940613,
+      "learning_rate": 5.753428561247415e-07,
+      "loss": -0.1153,
+      "num_tokens": 2204116.0,
+      "reward": 0.75,
+      "reward_std": 0.5,
+      "rewards/accuracy_reward": 0.75,
+      "step": 152
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1804.0,
+      "epoch": 0.003795113481334491,
+      "grad_norm": 0.0,
+      "learning_rate": 5.695865504800327e-07,
+      "loss": 0.0,
+      "num_tokens": 2212232.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 153
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 456.75,
+      "epoch": 0.003819918144611187,
+      "grad_norm": 0.0,
+      "learning_rate": 5.638208239302974e-07,
+      "loss": 0.0,
+      "num_tokens": 2214579.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 154
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5440.5,
+      "epoch": 0.003844722807887883,
+      "grad_norm": 0.0,
+      "learning_rate": 5.580464570626151e-07,
+      "loss": 0.0,
+      "num_tokens": 2236929.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 155
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1472.0,
+      "epoch": 0.003869527471164579,
+      "grad_norm": 0.0,
+      "learning_rate": 5.522642316338268e-07,
+      "loss": 0.0,
+      "num_tokens": 2243433.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 156
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1144.0,
+      "epoch": 0.003894332134441275,
+      "grad_norm": 0.0,
+      "learning_rate": 5.464749304646961e-07,
+      "loss": 0.0,
+      "num_tokens": 2248525.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 157
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2829.5,
+      "epoch": 0.003919136797717971,
+      "grad_norm": 0.0,
+      "learning_rate": 5.406793373339292e-07,
+      "loss": 0.0,
+      "num_tokens": 2260291.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 158
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7811.0,
+      "epoch": 0.003943941460994667,
+      "grad_norm": 0.0,
+      "learning_rate": 5.348782368720625e-07,
+      "loss": 0.0,
+      "num_tokens": 2292039.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 159
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3727.5,
+      "epoch": 0.003968746124271363,
+      "grad_norm": 0.0,
+      "learning_rate": 5.290724144552379e-07,
+      "loss": 0.0,
+      "num_tokens": 2307989.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 160
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3684.75,
+      "epoch": 0.003993550787548059,
+      "grad_norm": 0.0,
+      "learning_rate": 5.232626560988734e-07,
+      "loss": 0.0,
+      "num_tokens": 2323668.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 161
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6380.25,
+      "epoch": 0.004018355450824755,
+      "grad_norm": 0.0,
+      "learning_rate": 5.174497483512505e-07,
+      "loss": 0.0,
+      "num_tokens": 2350017.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 162
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7492.25,
+      "epoch": 0.004043160114101451,
+      "grad_norm": 0.0,
+      "learning_rate": 5.116344781870281e-07,
+      "loss": 0.0,
+      "num_tokens": 2380890.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 163
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4281.25,
+      "epoch": 0.004067964777378147,
+      "grad_norm": 0.0,
+      "learning_rate": 5.058176329006985e-07,
+      "loss": 0.0,
+      "num_tokens": 2398691.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 164
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5504.25,
+      "epoch": 0.004092769440654843,
+      "grad_norm": 0.0,
+      "learning_rate": 5e-07,
+      "loss": 0.0,
+      "num_tokens": 2421208.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 165
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2762.0,
+      "epoch": 0.004117574103931539,
+      "grad_norm": 0.0,
+      "learning_rate": 4.941823670993015e-07,
+      "loss": 0.0,
+      "num_tokens": 2432736.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 166
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3381.0,
+      "epoch": 0.004142378767208235,
+      "grad_norm": 0.0,
+      "learning_rate": 4.883655218129719e-07,
+      "loss": 0.0,
+      "num_tokens": 2446848.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 167
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3178.5,
+      "epoch": 0.004167183430484931,
+      "grad_norm": 0.0,
+      "learning_rate": 4.825502516487496e-07,
+      "loss": 0.0,
+      "num_tokens": 2460906.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 168
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2739.0,
+      "epoch": 0.004191988093761628,
+      "grad_norm": 0.0,
+      "learning_rate": 4.7673734390112666e-07,
+      "loss": 0.0,
+      "num_tokens": 2472598.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 169
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7138.25,
+      "epoch": 0.004216792757038323,
+      "grad_norm": 0.0,
+      "learning_rate": 4.7092758554476206e-07,
+      "loss": 0.0,
+      "num_tokens": 2501723.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 170
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7294.75,
+      "epoch": 0.0042415974203150195,
+      "grad_norm": 0.17870603501796722,
+      "learning_rate": 4.6512176312793735e-07,
+      "loss": -0.018,
+      "num_tokens": 2531502.0,
+      "reward": 0.25,
+      "reward_std": 0.5,
+      "rewards/accuracy_reward": 0.25,
+      "step": 171
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5977.25,
+      "epoch": 0.004266402083591715,
+      "grad_norm": 0.0,
+      "learning_rate": 4.593206626660709e-07,
+      "loss": 0.0,
+      "num_tokens": 2555995.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 172
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5944.5,
+      "epoch": 0.004291206746868411,
+      "grad_norm": 0.0,
+      "learning_rate": 4.535250695353039e-07,
+      "loss": 0.0,
+      "num_tokens": 2580333.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 173
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 974.75,
+      "epoch": 0.004316011410145107,
+      "grad_norm": 0.0,
+      "learning_rate": 4.477357683661733e-07,
+      "loss": 0.0,
+      "num_tokens": 2584660.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 174
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4904.25,
+      "epoch": 0.004340816073421803,
+      "grad_norm": 0.0,
+      "learning_rate": 4.419535429373848e-07,
+      "loss": 0.0,
+      "num_tokens": 2605065.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 175
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6226.0,
+      "epoch": 0.0043656207366985,
+      "grad_norm": 0.0,
+      "learning_rate": 4.3617917606970267e-07,
+      "loss": 0.0,
+      "num_tokens": 2630729.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 176
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 743.5,
+      "epoch": 0.004390425399975195,
+      "grad_norm": 0.0,
+      "learning_rate": 4.304134495199674e-07,
+      "loss": 0.0,
+      "num_tokens": 2634155.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 177
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1183.5,
+      "epoch": 0.004415230063251892,
+      "grad_norm": 0.0,
+      "learning_rate": 4.246571438752584e-07,
+      "loss": 0.0,
+      "num_tokens": 2639329.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 178
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6240.75,
+      "epoch": 0.004440034726528587,
+      "grad_norm": 0.0,
+      "learning_rate": 4.1891103844721634e-07,
+      "loss": 0.0,
+      "num_tokens": 2664792.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 179
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1472.0,
+      "epoch": 0.0044648393898052835,
+      "grad_norm": 0.0,
+      "learning_rate": 4.131759111665348e-07,
+      "loss": 0.0,
+      "num_tokens": 2671380.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 180
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3143.5,
+      "epoch": 0.004489644053081979,
+      "grad_norm": 0.2921382188796997,
+      "learning_rate": 4.074525384776428e-07,
+      "loss": 0.0159,
+      "num_tokens": 2684494.0,
+      "reward": 0.5,
+      "reward_std": 0.5773502588272095,
+      "rewards/accuracy_reward": 0.5,
+      "step": 181
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7197.5,
+      "epoch": 0.004514448716358675,
+      "grad_norm": 0.0,
+      "learning_rate": 4.0174169523358485e-07,
+      "loss": 0.0,
+      "num_tokens": 2713816.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 182
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2604.0,
+      "epoch": 0.004539253379635372,
+      "grad_norm": 0.0,
+      "learning_rate": 3.960441545911204e-07,
+      "loss": 0.0,
+      "num_tokens": 2725404.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 183
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5705.5,
+      "epoch": 0.004564058042912067,
+      "grad_norm": 0.0,
+      "learning_rate": 3.9036068790604823e-07,
+      "loss": 0.0,
+      "num_tokens": 2748802.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 184
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6358.75,
+      "epoch": 0.004588862706188764,
+      "grad_norm": 0.16501788794994354,
+      "learning_rate": 3.846920646287799e-07,
+      "loss": 0.0842,
+      "num_tokens": 2774653.0,
+      "reward": 0.25,
+      "reward_std": 0.5,
+      "rewards/accuracy_reward": 0.25,
+      "step": 185
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3738.0,
+      "epoch": 0.004613667369465459,
+      "grad_norm": 0.3154926300048828,
+      "learning_rate": 3.790390522001662e-07,
+      "loss": 0.0946,
+      "num_tokens": 2790353.0,
+      "reward": 0.75,
+      "reward_std": 0.5,
+      "rewards/accuracy_reward": 0.75,
+      "step": 186
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3408.25,
+      "epoch": 0.004638472032742156,
+      "grad_norm": 0.0,
+      "learning_rate": 3.734024159475991e-07,
+      "loss": 0.0,
+      "num_tokens": 2804462.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 187
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 850.0,
+      "epoch": 0.004663276696018851,
+      "grad_norm": 0.0,
+      "learning_rate": 3.6778291898139903e-07,
+      "loss": 0.0,
+      "num_tokens": 2808434.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 188
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2068.0,
+      "epoch": 0.0046880813592955475,
+      "grad_norm": 0.0,
+      "learning_rate": 3.621813220915004e-07,
+      "loss": 0.0,
+      "num_tokens": 2817286.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 189
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1111.0,
+      "epoch": 0.004712886022572244,
+      "grad_norm": 0.0,
+      "learning_rate": 3.56598383644455e-07,
+      "loss": 0.0,
+      "num_tokens": 2822274.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 190
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 8170.5,
+      "epoch": 0.004737690685848939,
+      "grad_norm": 0.0,
+      "learning_rate": 3.5103485948075894e-07,
+      "loss": 0.0,
+      "num_tokens": 2855644.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 191
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7154.75,
+      "epoch": 0.004762495349125636,
+      "grad_norm": 0.0,
+      "learning_rate": 3.454915028125263e-07,
+      "loss": 0.0,
+      "num_tokens": 2884907.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 192
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4736.75,
+      "epoch": 0.004787300012402331,
+      "grad_norm": 0.0,
+      "learning_rate": 3.3996906412151417e-07,
+      "loss": 0.0,
+      "num_tokens": 2904514.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 193
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 593.0,
+      "epoch": 0.004812104675679028,
+      "grad_norm": 0.0,
+      "learning_rate": 3.34468291057521e-07,
+      "loss": 0.0,
+      "num_tokens": 2907338.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 194
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4472.75,
+      "epoch": 0.004836909338955724,
+      "grad_norm": 0.0,
+      "learning_rate": 3.2898992833716563e-07,
+      "loss": 0.0,
+      "num_tokens": 2925641.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 195
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 590.0,
+      "epoch": 0.00486171400223242,
+      "grad_norm": 0.0,
+      "learning_rate": 3.235347176430656e-07,
+      "loss": 0.0,
+      "num_tokens": 2928537.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 196
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2825.25,
+      "epoch": 0.004886518665509116,
+      "grad_norm": 0.0,
+      "learning_rate": 3.181033975234244e-07,
+      "loss": 0.0,
+      "num_tokens": 2940418.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 197
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6512.0,
+      "epoch": 0.0049113233287858115,
+      "grad_norm": 0.0,
+      "learning_rate": 3.1269670329204393e-07,
+      "loss": 0.0,
+      "num_tokens": 2967054.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 198
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 608.5,
+      "epoch": 0.004936127992062508,
+      "grad_norm": 0.0,
+      "learning_rate": 3.073153669287759e-07,
+      "loss": 0.0,
+      "num_tokens": 2969848.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 199
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3488.25,
+      "epoch": 0.004960932655339203,
+      "grad_norm": 0.0,
+      "learning_rate": 3.0196011698042156e-07,
+      "loss": 0.0,
+      "num_tokens": 2984313.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 200
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 647.0,
+      "epoch": 0.0049857373186159,
+      "grad_norm": 0.0,
+      "learning_rate": 2.9663167846209996e-07,
+      "loss": 0.0,
+      "num_tokens": 2987433.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 201
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1120.5,
+      "epoch": 0.005010541981892596,
+      "grad_norm": 0.0,
+      "learning_rate": 2.9133077275909107e-07,
+      "loss": 0.0,
+      "num_tokens": 2992407.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 202
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1203.75,
+      "epoch": 0.005035346645169292,
+      "grad_norm": 0.0,
+      "learning_rate": 2.86058117529173e-07,
+      "loss": 0.0,
+      "num_tokens": 2997654.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 203
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 324.75,
+      "epoch": 0.005060151308445988,
+      "grad_norm": 0.0,
+      "learning_rate": 2.808144266054612e-07,
+      "loss": 0.0,
+      "num_tokens": 2999325.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 204
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7551.25,
+      "epoch": 0.005084955971722684,
+      "grad_norm": 0.0,
+      "learning_rate": 2.756004098997689e-07,
+      "loss": 0.0,
+      "num_tokens": 3029986.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 205
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4129.0,
+      "epoch": 0.00510976063499938,
+      "grad_norm": 0.0,
+      "learning_rate": 2.7041677330649406e-07,
+      "loss": 0.0,
+      "num_tokens": 3047106.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 206
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1562.5,
+      "epoch": 0.0051345652982760755,
+      "grad_norm": 0.0,
+      "learning_rate": 2.6526421860705473e-07,
+      "loss": 0.0,
+      "num_tokens": 3053824.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 207
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 8058.25,
+      "epoch": 0.005159369961552772,
+      "grad_norm": 0.0,
+      "learning_rate": 2.6014344337487703e-07,
+      "loss": 0.0,
+      "num_tokens": 3086921.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 208
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6682.0,
+      "epoch": 0.005184174624829468,
+      "grad_norm": 0.0,
+      "learning_rate": 2.550551408809565e-07,
+      "loss": 0.0,
+      "num_tokens": 3114257.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 209
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1586.25,
+      "epoch": 0.005208979288106164,
+      "grad_norm": 0.0,
+      "learning_rate": 2.500000000000001e-07,
+      "loss": 0.0,
+      "num_tokens": 3121206.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 210
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4050.5,
+      "epoch": 0.00523378395138286,
+      "grad_norm": 0.0,
+      "learning_rate": 2.4497870511716237e-07,
+      "loss": 0.0,
+      "num_tokens": 3138020.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 211
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1723.0,
+      "epoch": 0.005258588614659556,
+      "grad_norm": 0.0,
+      "learning_rate": 2.399919360353923e-07,
+      "loss": 0.0,
+      "num_tokens": 3145332.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 212
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3516.75,
+      "epoch": 0.005283393277936252,
+      "grad_norm": 0.0,
+      "learning_rate": 2.350403678833976e-07,
+      "loss": 0.0,
+      "num_tokens": 3160275.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 213
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3455.0,
+      "epoch": 0.0053081979412129485,
+      "grad_norm": 0.0,
+      "learning_rate": 2.301246710242437e-07,
+      "loss": 0.0,
+      "num_tokens": 3175927.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 214
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5161.5,
+      "epoch": 0.005333002604489644,
+      "grad_norm": 0.0,
+      "learning_rate": 2.25245510964597e-07,
+      "loss": 0.0,
+      "num_tokens": 3197301.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 215
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3201.25,
+      "epoch": 0.00535780726776634,
+      "grad_norm": 0.0,
+      "learning_rate": 2.2040354826462664e-07,
+      "loss": 0.0,
+      "num_tokens": 3210914.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 216
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4491.75,
+      "epoch": 0.005382611931043036,
+      "grad_norm": 0.0,
+      "learning_rate": 2.155994384485742e-07,
+      "loss": 0.0,
+      "num_tokens": 3229381.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 217
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2169.0,
+      "epoch": 0.005407416594319732,
+      "grad_norm": 0.0,
+      "learning_rate": 2.1083383191600672e-07,
+      "loss": 0.0,
+      "num_tokens": 3238517.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 218
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3591.0,
+      "epoch": 0.005432221257596428,
+      "grad_norm": 0.0,
+      "learning_rate": 2.0610737385376348e-07,
+      "loss": 0.0,
+      "num_tokens": 3253581.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 219
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2297.5,
+      "epoch": 0.005457025920873124,
+      "grad_norm": 0.0,
+      "learning_rate": 2.01420704148607e-07,
+      "loss": 0.0,
+      "num_tokens": 3263263.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 220
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5212.25,
+      "epoch": 0.0054818305841498206,
+      "grad_norm": 0.0,
+      "learning_rate": 1.9677445730059344e-07,
+      "loss": 0.0,
+      "num_tokens": 3284604.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 221
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 706.25,
+      "epoch": 0.005506635247426516,
+      "grad_norm": 0.0,
+      "learning_rate": 1.9216926233717084e-07,
+      "loss": 0.0,
+      "num_tokens": 3287817.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 222
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 545.5,
+      "epoch": 0.0055314399107032125,
+      "grad_norm": 0.0,
+      "learning_rate": 1.8760574272801998e-07,
+      "loss": 0.0,
+      "num_tokens": 3290371.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 223
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2282.0,
+      "epoch": 0.005556244573979908,
+      "grad_norm": 0.0,
+      "learning_rate": 1.830845163006448e-07,
+      "loss": 0.0,
+      "num_tokens": 3300159.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 224
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 825.75,
+      "epoch": 0.005581049237256604,
+      "grad_norm": 0.0,
+      "learning_rate": 1.7860619515673032e-07,
+      "loss": 0.0,
+      "num_tokens": 3303922.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 225
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 8192.0,
+      "epoch": 0.0056058539005333,
+      "grad_norm": 0.0,
+      "learning_rate": 1.741713855892724e-07,
+      "loss": 0.0,
+      "num_tokens": 3337458.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 226
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1880.75,
+      "epoch": 0.005630658563809996,
+      "grad_norm": 0.0,
+      "learning_rate": 1.697806880004962e-07,
+      "loss": 0.0,
+      "num_tokens": 3345573.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 227
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5774.0,
+      "epoch": 0.005655463227086693,
+      "grad_norm": 0.0,
+      "learning_rate": 1.6543469682057104e-07,
+      "loss": 0.0,
+      "num_tokens": 3369561.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 228
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1106.0,
+      "epoch": 0.005680267890363388,
+      "grad_norm": 0.0,
+      "learning_rate": 1.6113400042713388e-07,
+      "loss": 0.0,
+      "num_tokens": 3374385.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 229
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 625.0,
+      "epoch": 0.005705072553640085,
+      "grad_norm": 0.0,
+      "learning_rate": 1.5687918106563325e-07,
+      "loss": 0.0,
+      "num_tokens": 3377409.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 230
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 991.75,
+      "epoch": 0.00572987721691678,
+      "grad_norm": 0.0,
+      "learning_rate": 1.5267081477050131e-07,
+      "loss": 0.0,
+      "num_tokens": 3382456.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 231
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 8192.0,
+      "epoch": 0.0057546818801934765,
+      "grad_norm": 0.0,
+      "learning_rate": 1.4850947128716911e-07,
+      "loss": 0.0,
+      "num_tokens": 3416072.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 232
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3301.25,
+      "epoch": 0.005779486543470172,
+      "grad_norm": 0.0,
+      "learning_rate": 1.4439571399493145e-07,
+      "loss": 0.0,
+      "num_tokens": 3430053.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 233
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2911.5,
+      "epoch": 0.005804291206746868,
+      "grad_norm": 0.0,
+      "learning_rate": 1.4033009983067452e-07,
+      "loss": 0.0,
+      "num_tokens": 3442327.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 234
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3780.0,
+      "epoch": 0.005829095870023565,
+      "grad_norm": 0.0,
+      "learning_rate": 1.3631317921347562e-07,
+      "loss": 0.0,
+      "num_tokens": 3458067.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 235
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3959.25,
+      "epoch": 0.00585390053330026,
+      "grad_norm": 0.0,
+      "learning_rate": 1.3234549597008572e-07,
+      "loss": 0.0,
+      "num_tokens": 3475240.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 236
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1576.75,
+      "epoch": 0.005878705196576957,
+      "grad_norm": 0.0,
+      "learning_rate": 1.284275872613028e-07,
+      "loss": 0.0,
+      "num_tokens": 3482159.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 237
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2267.5,
+      "epoch": 0.005903509859853652,
+      "grad_norm": 0.0,
+      "learning_rate": 1.245599835092504e-07,
+      "loss": 0.0,
+      "num_tokens": 3491753.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 238
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6582.25,
+      "epoch": 0.005928314523130349,
+      "grad_norm": 0.0,
+      "learning_rate": 1.2074320832556556e-07,
+      "loss": 0.0,
+      "num_tokens": 3518766.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 239
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4409.5,
+      "epoch": 0.005953119186407045,
+      "grad_norm": 0.0,
+      "learning_rate": 1.1697777844051104e-07,
+      "loss": 0.0,
+      "num_tokens": 3537124.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 240
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 744.75,
+      "epoch": 0.0059779238496837405,
+      "grad_norm": 0.0,
+      "learning_rate": 1.1326420363301808e-07,
+      "loss": 0.0,
+      "num_tokens": 3540559.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 241
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1890.0,
+      "epoch": 0.006002728512960437,
+      "grad_norm": 0.0,
+      "learning_rate": 1.096029866616704e-07,
+      "loss": 0.0,
+      "num_tokens": 3549251.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 242
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1161.75,
+      "epoch": 0.006027533176237132,
+      "grad_norm": 0.0,
+      "learning_rate": 1.0599462319663904e-07,
+      "loss": 0.0,
+      "num_tokens": 3554662.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 243
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2727.0,
+      "epoch": 0.006052337839513829,
+      "grad_norm": 0.0,
+      "learning_rate": 1.0243960175257605e-07,
+      "loss": 0.0,
+      "num_tokens": 3566014.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 244
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1661.25,
+      "epoch": 0.006077142502790524,
+      "grad_norm": 0.0,
+      "learning_rate": 9.893840362247807e-08,
+      "loss": 0.0,
+      "num_tokens": 3574155.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 245
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3048.5,
+      "epoch": 0.006101947166067221,
+      "grad_norm": 0.0,
+      "learning_rate": 9.549150281252632e-08,
+      "loss": 0.0,
+      "num_tokens": 3586801.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 246
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4632.0,
+      "epoch": 0.006126751829343917,
+      "grad_norm": 0.0,
+      "learning_rate": 9.209936597791407e-08,
+      "loss": 0.0,
+      "num_tokens": 3606109.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 247
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1765.5,
+      "epoch": 0.006151556492620613,
+      "grad_norm": 0.0,
+      "learning_rate": 8.876245235966883e-08,
+      "loss": 0.0,
+      "num_tokens": 3613815.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 248
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4083.75,
+      "epoch": 0.006176361155897309,
+      "grad_norm": 0.0,
+      "learning_rate": 8.548121372247919e-08,
+      "loss": 0.0,
+      "num_tokens": 3630678.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 249
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4663.75,
+      "epoch": 0.0062011658191740045,
+      "grad_norm": 0.0,
+      "learning_rate": 8.225609429353186e-08,
+      "loss": 0.0,
+      "num_tokens": 3649817.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 250
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2062.0,
+      "epoch": 0.006225970482450701,
+      "grad_norm": 0.0,
+      "learning_rate": 7.908753070237124e-08,
+      "loss": 0.0,
+      "num_tokens": 3658493.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 251
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1659.25,
+      "epoch": 0.006250775145727396,
+      "grad_norm": 0.0,
+      "learning_rate": 7.597595192178702e-08,
+      "loss": 0.0,
+      "num_tokens": 3665746.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 252
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1894.0,
+      "epoch": 0.006275579809004093,
+      "grad_norm": 0.0,
+      "learning_rate": 7.292177920973724e-08,
+      "loss": 0.0,
+      "num_tokens": 3673830.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 253
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5424.25,
+      "epoch": 0.006300384472280789,
+      "grad_norm": 0.0,
+      "learning_rate": 6.992542605231738e-08,
+      "loss": 0.0,
+      "num_tokens": 3695967.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 254
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 498.25,
+      "epoch": 0.006325189135557485,
+      "grad_norm": 0.0,
+      "learning_rate": 6.698729810778064e-08,
+      "loss": 0.0,
+      "num_tokens": 3698604.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 255
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 8192.0,
+      "epoch": 0.006349993798834181,
+      "grad_norm": 0.0,
+      "learning_rate": 6.410779315161886e-08,
+      "loss": 0.0,
+      "num_tokens": 3732176.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 256
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3048.0,
+      "epoch": 0.006374798462110877,
+      "grad_norm": 0.0,
+      "learning_rate": 6.128730102270896e-08,
+      "loss": 0.0,
+      "num_tokens": 3745092.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 257
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2695.25,
+      "epoch": 0.006399603125387573,
+      "grad_norm": 0.0,
+      "learning_rate": 5.8526203570536504e-08,
+      "loss": 0.0,
+      "num_tokens": 3756593.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 258
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 845.75,
+      "epoch": 0.0064244077886642685,
+      "grad_norm": 0.0,
+      "learning_rate": 5.582487460349805e-08,
+      "loss": 0.0,
+      "num_tokens": 3760404.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 259
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2017.75,
+      "epoch": 0.006449212451940965,
+      "grad_norm": 0.0,
+      "learning_rate": 5.318367983829392e-08,
+      "loss": 0.0,
+      "num_tokens": 3769003.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 260
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 8191.5,
+      "epoch": 0.006474017115217661,
+      "grad_norm": 0.0,
+      "learning_rate": 5.060297685041659e-08,
+      "loss": 0.0,
+      "num_tokens": 3802305.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 261
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1256.5,
+      "epoch": 0.006498821778494357,
+      "grad_norm": 0.0,
+      "learning_rate": 4.808311502573975e-08,
+      "loss": 0.0,
+      "num_tokens": 3808003.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 262
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7104.5,
+      "epoch": 0.006523626441771053,
+      "grad_norm": 0.0,
+      "learning_rate": 4.5624435513217873e-08,
+      "loss": 0.0,
+      "num_tokens": 3837061.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 263
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3343.25,
+      "epoch": 0.006548431105047749,
+      "grad_norm": 0.0,
+      "learning_rate": 4.322727117869951e-08,
+      "loss": 0.0,
+      "num_tokens": 3851114.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 264
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5981.75,
+      "epoch": 0.006573235768324445,
+      "grad_norm": 0.20941807329654694,
+      "learning_rate": 4.089194655986306e-08,
+      "loss": 0.1413,
+      "num_tokens": 3875585.0,
+      "reward": 0.25,
+      "reward_std": 0.5,
+      "rewards/accuracy_reward": 0.25,
+      "step": 265
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2008.5,
+      "epoch": 0.0065980404316011414,
+      "grad_norm": 0.0,
+      "learning_rate": 3.861877782227885e-08,
+      "loss": 0.0,
+      "num_tokens": 3884311.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 266
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6928.75,
+      "epoch": 0.006622845094877837,
+      "grad_norm": 0.0,
+      "learning_rate": 3.6408072716606345e-08,
+      "loss": 0.0,
+      "num_tokens": 3912594.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 267
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 4809.0,
+      "epoch": 0.006647649758154533,
+      "grad_norm": 0.0,
+      "learning_rate": 3.426013053692878e-08,
+      "loss": 0.0,
+      "num_tokens": 3932422.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 268
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6333.5,
+      "epoch": 0.006672454421431229,
+      "grad_norm": 0.0,
+      "learning_rate": 3.217524208023431e-08,
+      "loss": 0.0,
+      "num_tokens": 3958284.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 269
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2418.75,
+      "epoch": 0.006697259084707925,
+      "grad_norm": 0.0,
+      "learning_rate": 3.015368960704584e-08,
+      "loss": 0.0,
+      "num_tokens": 3968403.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 270
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6057.25,
+      "epoch": 0.006722063747984621,
+      "grad_norm": 0.0,
+      "learning_rate": 2.8195746803208242e-08,
+      "loss": 0.0,
+      "num_tokens": 3993200.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 271
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 821.5,
+      "epoch": 0.006746868411261317,
+      "grad_norm": 0.0,
+      "learning_rate": 2.6301678742835397e-08,
+      "loss": 0.0,
+      "num_tokens": 3996950.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 272
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 8192.0,
+      "epoch": 0.0067716730745380135,
+      "grad_norm": 0.0,
+      "learning_rate": 2.4471741852423233e-08,
+      "loss": 0.0,
+      "num_tokens": 4030314.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 273
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2693.5,
+      "epoch": 0.006796477737814709,
+      "grad_norm": 0.0,
+      "learning_rate": 2.2706183876134045e-08,
+      "loss": 0.0,
+      "num_tokens": 4041892.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 274
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 965.25,
+      "epoch": 0.0068212824010914054,
+      "grad_norm": 0.0,
+      "learning_rate": 2.100524384225555e-08,
+      "loss": 0.0,
+      "num_tokens": 4046237.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 275
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3427.25,
+      "epoch": 0.006846087064368101,
+      "grad_norm": 0.0,
+      "learning_rate": 1.936915203084055e-08,
+      "loss": 0.0,
+      "num_tokens": 4060802.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 276
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 7794.75,
+      "epoch": 0.006870891727644797,
+      "grad_norm": 0.0,
+      "learning_rate": 1.7798129942530548e-08,
+      "loss": 0.0,
+      "num_tokens": 4092813.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 277
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6520.0,
+      "epoch": 0.006895696390921493,
+      "grad_norm": 0.0,
+      "learning_rate": 1.62923902685681e-08,
+      "loss": 0.0,
+      "num_tokens": 4119641.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 278
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1263.0,
+      "epoch": 0.006920501054198189,
+      "grad_norm": 0.0,
+      "learning_rate": 1.4852136862001763e-08,
+      "loss": 0.0,
+      "num_tokens": 4125333.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 279
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1784.0,
+      "epoch": 0.006945305717474886,
+      "grad_norm": 0.0,
+      "learning_rate": 1.3477564710088096e-08,
+      "loss": 0.0,
+      "num_tokens": 4132957.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 280
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 5096.0,
+      "epoch": 0.006970110380751581,
+      "grad_norm": 0.0,
+      "learning_rate": 1.2168859907892902e-08,
+      "loss": 0.0,
+      "num_tokens": 4154117.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 281
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1427.0,
+      "epoch": 0.0069949150440282775,
+      "grad_norm": 0.0,
+      "learning_rate": 1.0926199633097154e-08,
+      "loss": 0.0,
+      "num_tokens": 4160381.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 282
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 862.25,
+      "epoch": 0.007019719707304973,
+      "grad_norm": 0.0,
+      "learning_rate": 9.749752122010346e-09,
+      "loss": 0.0,
+      "num_tokens": 4164318.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 283
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6548.25,
+      "epoch": 0.0070445243705816694,
+      "grad_norm": 0.0,
+      "learning_rate": 8.639676646793382e-09,
+      "loss": 0.0,
+      "num_tokens": 4191083.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 284
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1828.5,
+      "epoch": 0.007069329033858365,
+      "grad_norm": 0.0,
+      "learning_rate": 7.59612349389599e-09,
+      "loss": 0.0,
+      "num_tokens": 4198809.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 285
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1426.0,
+      "epoch": 0.007094133697135061,
+      "grad_norm": 0.405701607465744,
+      "learning_rate": 6.61923394371039e-09,
+      "loss": 0.0284,
+      "num_tokens": 4204961.0,
+      "reward": 0.75,
+      "reward_std": 0.5,
+      "rewards/accuracy_reward": 0.75,
+      "step": 286
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6286.25,
+      "epoch": 0.007118938360411758,
+      "grad_norm": 0.0,
+      "learning_rate": 5.7091402514442e-09,
+      "loss": 0.0,
+      "num_tokens": 4230786.0,
+      "reward": 0.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 0.0,
+      "step": 287
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1944.25,
+      "epoch": 0.007143743023688453,
+      "grad_norm": 0.5905258655548096,
+      "learning_rate": 4.865965629214819e-09,
+      "loss": -0.042,
+      "num_tokens": 4239071.0,
+      "reward": 0.25,
+      "reward_std": 0.5,
+      "rewards/accuracy_reward": 0.25,
+      "step": 288
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 860.5,
+      "epoch": 0.00716854768696515,
+      "grad_norm": 0.0,
+      "learning_rate": 4.089824229369154e-09,
+      "loss": 0.0,
+      "num_tokens": 4243021.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 289
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 669.5,
+      "epoch": 0.007193352350241845,
+      "grad_norm": 0.0,
+      "learning_rate": 3.380821129028488e-09,
+      "loss": 0.0,
+      "num_tokens": 4246303.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 290
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1785.75,
+      "epoch": 0.0072181570135185415,
+      "grad_norm": 0.0,
+      "learning_rate": 2.739052315863355e-09,
+      "loss": 0.0,
+      "num_tokens": 4254082.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 291
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 3101.5,
+      "epoch": 0.007242961676795238,
+      "grad_norm": 0.0,
+      "learning_rate": 2.1646046750978253e-09,
+      "loss": 0.0,
+      "num_tokens": 4266976.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 292
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 608.25,
+      "epoch": 0.0072677663400719334,
+      "grad_norm": 0.0,
+      "learning_rate": 1.6575559777469717e-09,
+      "loss": 0.0,
+      "num_tokens": 4269801.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 293
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 6663.0,
+      "epoch": 0.00729257100334863,
+      "grad_norm": 0.0,
+      "learning_rate": 1.217974870087901e-09,
+      "loss": 0.0,
+      "num_tokens": 4297549.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 294
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 803.75,
+      "epoch": 0.007317375666625325,
+      "grad_norm": 0.0,
+      "learning_rate": 8.459208643659121e-10,
+      "loss": 0.0,
+      "num_tokens": 4301396.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 295
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1361.5,
+      "epoch": 0.007342180329902022,
+      "grad_norm": 0.0,
+      "learning_rate": 5.41444330737717e-10,
+      "loss": 0.0,
+      "num_tokens": 4307358.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 296
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 532.5,
+      "epoch": 0.007366984993178717,
+      "grad_norm": 0.0,
+      "learning_rate": 3.0458649045211894e-10,
+      "loss": 0.0,
+      "num_tokens": 4309896.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 297
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 1515.5,
+      "epoch": 0.007391789656455414,
+      "grad_norm": 0.0,
+      "learning_rate": 1.3537941026914302e-10,
+      "loss": 0.0,
+      "num_tokens": 4316658.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 298
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 8192.0,
+      "epoch": 0.00741659431973211,
+      "grad_norm": 0.0,
+      "learning_rate": 3.3845998118897657e-11,
+      "loss": 0.0,
+      "num_tokens": 4350074.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 299
+    },
+    {
+      "clip_ratio": 0.0,
+      "completion_length": 2304.25,
+      "epoch": 0.0074413989830088055,
+      "grad_norm": 0.0,
+      "learning_rate": 0.0,
+      "loss": 0.0,
+      "num_tokens": 4359795.0,
+      "reward": 1.0,
+      "reward_std": 0.0,
+      "rewards/accuracy_reward": 1.0,
+      "step": 300
+    },
+    {
+      "epoch": 0.0074413989830088055,
+      "step": 300,
+      "total_flos": 0.0,
+      "train_loss": 0.001285893005115213,
+      "train_runtime": 7938.9492,
+      "train_samples_per_second": 0.151,
+      "train_steps_per_second": 0.038
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 300,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 1,
+  "save_steps": 50,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 4,
+  "trial_name": null,
+  "trial_params": null
+}