diff --git "a/checkpoint-3000/trainer_state.json" "b/checkpoint-3000/trainer_state.json"
new file mode 100644--- /dev/null
+++ "b/checkpoint-3000/trainer_state.json"
@@ -0,0 +1,27033 @@
+{
+  "best_metric": null,
+  "best_model_checkpoint": null,
+  "epoch": 1.6016,
+  "eval_steps": 500,
+  "global_step": 3000,
+  "is_hyper_param_search": false,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "completion_length": 253.59375,
+      "epoch": 0.0005333333333333334,
+      "grad_norm": 0.1840473860502243,
+      "learning_rate": 3e-06,
+      "loss": 0.0686,
+      "policy/loss": -0.013315461575984955,
+      "reward": 1.3489583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.640625,
+      "rewards/correctness_reward_func_math": 0.7083333283662796,
+      "step": 1
+    },
+    {
+      "epoch": 0.0010666666666666667,
+      "grad_norm": 0.21767450869083405,
+      "learning_rate": 3e-06,
+      "loss": -0.0802,
+      "policy/loss": 0.0014063939452171326,
+      "step": 2
+    },
+    {
+      "epoch": 0.0016,
+      "grad_norm": 0.18407362699508667,
+      "learning_rate": 3e-06,
+      "loss": 0.0688,
+      "policy/loss": -0.011666037142276764,
+      "step": 3
+    },
+    {
+      "epoch": 0.0021333333333333334,
+      "grad_norm": 0.25259828567504883,
+      "learning_rate": 3e-06,
+      "loss": -0.0806,
+      "policy/loss": 0.002615857869386673,
+      "step": 4
+    },
+    {
+      "completion_length": 229.6666717529297,
+      "epoch": 0.0026666666666666666,
+      "grad_norm": 0.1706053763628006,
+      "learning_rate": 3e-06,
+      "loss": 0.118,
+      "policy/loss": 0.13203176095328217,
+      "reward": 1.2500000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 5
+    },
+    {
+      "epoch": 0.0032,
+      "grad_norm": 0.16035571694374084,
+      "learning_rate": 3e-06,
+      "loss": -0.0034,
+      "policy/loss": -0.01066783096733559,
+      "step": 6
+    },
+    {
+      "epoch": 0.0037333333333333333,
+      "grad_norm": 0.18553981184959412,
+      "learning_rate": 3e-06,
+      "loss": 0.1178,
+      "policy/loss": 0.1316425723646404,
+      "step": 7
+    },
+    {
+      "epoch": 0.004266666666666667,
+      "grad_norm": 0.15817661583423615,
+      "learning_rate": 3e-06,
+      "loss": -0.0036,
+      "policy/loss": -0.010556348789236836,
+      "step": 8
+    },
+    {
+      "completion_length": 246.15625762939453,
+      "epoch": 0.0048,
+      "grad_norm": 0.17941085994243622,
+      "learning_rate": 3e-06,
+      "loss": 0.1236,
+      "policy/loss": 0.1340911271981895,
+      "reward": 1.1718750298023224,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5677083283662796,
+      "rewards/correctness_reward_func_math": 0.6041666567325592,
+      "step": 9
+    },
+    {
+      "epoch": 0.005333333333333333,
+      "grad_norm": 0.18234595656394958,
+      "learning_rate": 3e-06,
+      "loss": -0.1056,
+      "policy/loss": -0.6010678485035896,
+      "step": 10
+    },
+    {
+      "epoch": 0.005866666666666667,
+      "grad_norm": 0.18676400184631348,
+      "learning_rate": 3e-06,
+      "loss": 0.1234,
+      "policy/loss": 0.13190458621829748,
+      "step": 11
+    },
+    {
+      "epoch": 0.0064,
+      "grad_norm": 0.1809026598930359,
+      "learning_rate": 3e-06,
+      "loss": -0.1059,
+      "policy/loss": -0.6034602895379066,
+      "step": 12
+    },
+    {
+      "completion_length": 232.28125762939453,
+      "epoch": 0.006933333333333333,
+      "grad_norm": 0.23839235305786133,
+      "learning_rate": 3e-06,
+      "loss": 0.0539,
+      "policy/loss": 0.217545285820961,
+      "reward": 1.0937500298023224,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.4375000149011612,
+      "step": 13
+    },
+    {
+      "epoch": 0.007466666666666667,
+      "grad_norm": 0.20470507442951202,
+      "learning_rate": 3e-06,
+      "loss": 0.1624,
+      "policy/loss": 0.3049881956540048,
+      "step": 14
+    },
+    {
+      "epoch": 0.008,
+      "grad_norm": 0.23264506459236145,
+      "learning_rate": 3e-06,
+      "loss": 0.052,
+      "policy/loss": 0.21768283285200596,
+      "step": 15
+    },
+    {
+      "epoch": 0.008533333333333334,
+      "grad_norm": 0.22143056988716125,
+      "learning_rate": 3e-06,
+      "loss": 0.1609,
+      "policy/loss": 0.3069990575313568,
+      "step": 16
+    },
+    {
+      "completion_length": 248.75000762939453,
+      "epoch": 0.009066666666666667,
+      "grad_norm": 0.16627700626850128,
+      "learning_rate": 3e-06,
+      "loss": 0.1356,
+      "policy/loss": 0.26722581312060356,
+      "reward": 1.15625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.59375,
+      "rewards/correctness_reward_func_math": 0.5624999850988388,
+      "step": 17
+    },
+    {
+      "epoch": 0.0096,
+      "grad_norm": 0.24618378281593323,
+      "learning_rate": 3e-06,
+      "loss": 0.1849,
+      "policy/loss": -0.019017666578292847,
+      "step": 18
+    },
+    {
+      "epoch": 0.010133333333333333,
+      "grad_norm": 0.17481371760368347,
+      "learning_rate": 3e-06,
+      "loss": 0.1363,
+      "policy/loss": 0.2626461684703827,
+      "step": 19
+    },
+    {
+      "epoch": 0.010666666666666666,
+      "grad_norm": 0.2184562087059021,
+      "learning_rate": 3e-06,
+      "loss": 0.1838,
+      "policy/loss": -0.019053876399993896,
+      "step": 20
+    },
+    {
+      "completion_length": 244.75,
+      "epoch": 0.0112,
+      "grad_norm": 0.1663179248571396,
+      "learning_rate": 3e-06,
+      "loss": 0.0782,
+      "policy/loss": 0.1924348697066307,
+      "reward": 1.1927083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.609375,
+      "rewards/correctness_reward_func_math": 0.5833333432674408,
+      "step": 21
+    },
+    {
+      "epoch": 0.011733333333333333,
+      "grad_norm": 0.2521425783634186,
+      "learning_rate": 3e-06,
+      "loss": 0.0032,
+      "policy/loss": 0.272992268204689,
+      "step": 22
+    },
+    {
+      "epoch": 0.012266666666666667,
+      "grad_norm": 0.17646797001361847,
+      "learning_rate": 3e-06,
+      "loss": 0.0778,
+      "policy/loss": 0.1907222494482994,
+      "step": 23
+    },
+    {
+      "epoch": 0.0128,
+      "grad_norm": 0.2792782783508301,
+      "learning_rate": 3e-06,
+      "loss": 0.0038,
+      "policy/loss": 0.2760181352496147,
+      "step": 24
+    },
+    {
+      "completion_length": 239.31250762939453,
+      "epoch": 0.013333333333333334,
+      "grad_norm": 0.1828080415725708,
+      "learning_rate": 3e-06,
+      "loss": -0.1027,
+      "policy/loss": 0.14645338617265224,
+      "reward": 1.1875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5833333134651184,
+      "rewards/correctness_reward_func_math": 0.6041666716337204,
+      "step": 25
+    },
+    {
+      "epoch": 0.013866666666666666,
+      "grad_norm": 0.23987236618995667,
+      "learning_rate": 3e-06,
+      "loss": 0.0097,
+      "policy/loss": 0.023717982694506645,
+      "step": 26
+    },
+    {
+      "epoch": 0.0144,
+      "grad_norm": 0.17299237847328186,
+      "learning_rate": 3e-06,
+      "loss": -0.1038,
+      "policy/loss": 0.14596972800791264,
+      "step": 27
+    },
+    {
+      "epoch": 0.014933333333333333,
+      "grad_norm": 0.17266596853733063,
+      "learning_rate": 3e-06,
+      "loss": 0.0096,
+      "policy/loss": 0.024075211957097054,
+      "step": 28
+    },
+    {
+      "completion_length": 247.625,
+      "epoch": 0.015466666666666667,
+      "grad_norm": 0.2530646324157715,
+      "learning_rate": 3e-06,
+      "loss": -0.0041,
+      "policy/loss": -0.05582362553104758,
+      "reward": 1.0000000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5833333432674408,
+      "rewards/correctness_reward_func_math": 0.4166666716337204,
+      "step": 29
+    },
+    {
+      "epoch": 0.016,
+      "grad_norm": 0.12944753468036652,
+      "learning_rate": 3e-06,
+      "loss": 0.0217,
+      "policy/loss": 0.22346153855323792,
+      "step": 30
+    },
+    {
+      "epoch": 0.016533333333333334,
+      "grad_norm": 0.16591361165046692,
+      "learning_rate": 3e-06,
+      "loss": -0.004,
+      "policy/loss": -0.05613691732287407,
+      "step": 31
+    },
+    {
+      "epoch": 0.017066666666666667,
+      "grad_norm": 0.1316152960062027,
+      "learning_rate": 3e-06,
+      "loss": 0.0211,
+      "policy/loss": 0.22320326673798263,
+      "step": 32
+    },
+    {
+      "completion_length": 242.23959350585938,
+      "epoch": 0.0176,
+      "grad_norm": 0.1701315939426422,
+      "learning_rate": 3e-06,
+      "loss": -0.0747,
+      "policy/loss": -0.06686947916568897,
+      "reward": 0.9531250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.578125,
+      "rewards/correctness_reward_func_math": 0.375,
+      "step": 33
+    },
+    {
+      "epoch": 0.018133333333333335,
+      "grad_norm": 0.25609517097473145,
+      "learning_rate": 3e-06,
+      "loss": -0.0632,
+      "policy/loss": 0.6440966994451998,
+      "step": 34
+    },
+    {
+      "epoch": 0.018666666666666668,
+      "grad_norm": 0.1868312656879425,
+      "learning_rate": 3e-06,
+      "loss": -0.0743,
+      "policy/loss": -0.06481578611324323,
+      "step": 35
+    },
+    {
+      "epoch": 0.0192,
+      "grad_norm": 0.30514949560165405,
+      "learning_rate": 3e-06,
+      "loss": -0.0649,
+      "policy/loss": 0.6458015113778259,
+      "step": 36
+    },
+    {
+      "completion_length": 247.70833587646484,
+      "epoch": 0.019733333333333332,
+      "grad_norm": 0.19562242925167084,
+      "learning_rate": 3e-06,
+      "loss": -0.0029,
+      "policy/loss": 0.48862760227273583,
+      "reward": 1.026041716337204,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6302083134651184,
+      "rewards/correctness_reward_func_math": 0.3958333432674408,
+      "step": 37
+    },
+    {
+      "epoch": 0.020266666666666665,
+      "grad_norm": 0.1894940733909607,
+      "learning_rate": 3e-06,
+      "loss": -0.0825,
+      "policy/loss": -0.1527944498995737,
+      "step": 38
+    },
+    {
+      "epoch": 0.0208,
+      "grad_norm": 0.15389592945575714,
+      "learning_rate": 3e-06,
+      "loss": -0.0035,
+      "policy/loss": 0.488402475512288,
+      "step": 39
+    },
+    {
+      "epoch": 0.021333333333333333,
+      "grad_norm": 0.17074385285377502,
+      "learning_rate": 3e-06,
+      "loss": -0.0817,
+      "policy/loss": -0.14642204285543414,
+      "step": 40
+    },
+    {
+      "completion_length": 240.42708587646484,
+      "epoch": 0.021866666666666666,
+      "grad_norm": 0.1479763686656952,
+      "learning_rate": 3e-06,
+      "loss": -0.1517,
+      "policy/loss": -0.04031290794792852,
+      "reward": 1.0208333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.3958333283662796,
+      "step": 41
+    },
+    {
+      "epoch": 0.0224,
+      "grad_norm": 0.16938340663909912,
+      "learning_rate": 3e-06,
+      "loss": -0.0324,
+      "policy/loss": 0.22682796608215483,
+      "step": 42
+    },
+    {
+      "epoch": 0.022933333333333333,
+      "grad_norm": 0.20570379495620728,
+      "learning_rate": 3e-06,
+      "loss": -0.1513,
+      "policy/loss": -0.03991006245014006,
+      "step": 43
+    },
+    {
+      "epoch": 0.023466666666666667,
+      "grad_norm": 0.16357482969760895,
+      "learning_rate": 3e-06,
+      "loss": -0.0315,
+      "policy/loss": 0.22647795266613713,
+      "step": 44
+    },
+    {
+      "completion_length": 243.75000762939453,
+      "epoch": 0.024,
+      "grad_norm": 0.2123340517282486,
+      "learning_rate": 3e-06,
+      "loss": -0.0118,
+      "policy/loss": -0.5904987212270498,
+      "reward": 1.2604167461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6354166567325592,
+      "rewards/correctness_reward_func_math": 0.6250000149011612,
+      "step": 45
+    },
+    {
+      "epoch": 0.024533333333333334,
+      "grad_norm": 0.17777524888515472,
+      "learning_rate": 3e-06,
+      "loss": 0.0533,
+      "policy/loss": -0.11857842281460762,
+      "step": 46
+    },
+    {
+      "epoch": 0.025066666666666668,
+      "grad_norm": 0.19381994009017944,
+      "learning_rate": 3e-06,
+      "loss": -0.011,
+      "policy/loss": -0.5887860413640738,
+      "step": 47
+    },
+    {
+      "epoch": 0.0256,
+      "grad_norm": 0.17782829701900482,
+      "learning_rate": 3e-06,
+      "loss": 0.0535,
+      "policy/loss": -0.11902743950486183,
+      "step": 48
+    },
+    {
+      "completion_length": 249.58333587646484,
+      "epoch": 0.026133333333333335,
+      "grad_norm": 0.19484445452690125,
+      "learning_rate": 3e-06,
+      "loss": 0.0211,
+      "policy/loss": 0.10960811631071721,
+      "reward": 1.119791716337204,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5989583432674408,
+      "rewards/correctness_reward_func_math": 0.5208333432674408,
+      "step": 49
+    },
+    {
+      "epoch": 0.02666666666666667,
+      "grad_norm": 0.19596777856349945,
+      "learning_rate": 3e-06,
+      "loss": -0.0276,
+      "policy/loss": -0.2616497438689649,
+      "step": 50
+    },
+    {
+      "epoch": 0.0272,
+      "grad_norm": 0.1741992086172104,
+      "learning_rate": 3e-06,
+      "loss": 0.0208,
+      "policy/loss": 0.109885911165577,
+      "step": 51
+    },
+    {
+      "epoch": 0.027733333333333332,
+      "grad_norm": 0.20924681425094604,
+      "learning_rate": 3e-06,
+      "loss": -0.028,
+      "policy/loss": -0.26201973971166836,
+      "step": 52
+    },
+    {
+      "completion_length": 248.39583587646484,
+      "epoch": 0.028266666666666666,
+      "grad_norm": 0.20535294711589813,
+      "learning_rate": 3e-06,
+      "loss": 0.0115,
+      "policy/loss": 2.324437957668124e-07,
+      "reward": 1.2395833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5520833134651184,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 53
+    },
+    {
+      "epoch": 0.0288,
+      "grad_norm": 0.21366165578365326,
+      "learning_rate": 3e-06,
+      "loss": 0.0282,
+      "policy/loss": 2.0069056461125e-07,
+      "step": 54
+    },
+    {
+      "epoch": 0.029333333333333333,
+      "grad_norm": 0.19997094571590424,
+      "learning_rate": 3e-06,
+      "loss": 0.0101,
+      "policy/loss": 2.325854069340494e-07,
+      "step": 55
+    },
+    {
+      "epoch": 0.029866666666666666,
+      "grad_norm": 0.24604934453964233,
+      "learning_rate": 3e-06,
+      "loss": 0.0277,
+      "policy/loss": 2.0050400451054884e-07,
+      "step": 56
+    },
+    {
+      "completion_length": 244.12500762939453,
+      "epoch": 0.0304,
+      "grad_norm": 0.1484808325767517,
+      "learning_rate": 3e-06,
+      "loss": 0.0043,
+      "policy/loss": 0.1830545961856842,
+      "reward": 1.4114583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5989583432674408,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 57
+    },
+    {
+      "epoch": 0.030933333333333334,
+      "grad_norm": 0.16552947461605072,
+      "learning_rate": 3e-06,
+      "loss": 0.2109,
+      "policy/loss": 0.07458048802800477,
+      "step": 58
+    },
+    {
+      "epoch": 0.031466666666666664,
+      "grad_norm": 0.15123622119426727,
+      "learning_rate": 3e-06,
+      "loss": 0.0039,
+      "policy/loss": 0.18247901275753975,
+      "step": 59
+    },
+    {
+      "epoch": 0.032,
+      "grad_norm": 0.1817777156829834,
+      "learning_rate": 3e-06,
+      "loss": 0.2107,
+      "policy/loss": 0.07392013329081237,
+      "step": 60
+    },
+    {
+      "completion_length": 251.6666717529297,
+      "epoch": 0.03253333333333333,
+      "grad_norm": 0.15461507439613342,
+      "learning_rate": 3e-06,
+      "loss": -0.0304,
+      "policy/loss": -0.014731228351593018,
+      "reward": 0.9479166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.59375,
+      "rewards/correctness_reward_func_math": 0.3541666716337204,
+      "step": 61
+    },
+    {
+      "epoch": 0.03306666666666667,
+      "grad_norm": 0.16346685588359833,
+      "learning_rate": 3e-06,
+      "loss": 0.0217,
+      "policy/loss": 0.29023469239473343,
+      "step": 62
+    },
+    {
+      "epoch": 0.0336,
+      "grad_norm": 0.14802707731723785,
+      "learning_rate": 3e-06,
+      "loss": -0.0309,
+      "policy/loss": -0.014934249222278595,
+      "step": 63
+    },
+    {
+      "epoch": 0.034133333333333335,
+      "grad_norm": 0.15431180596351624,
+      "learning_rate": 3e-06,
+      "loss": 0.0201,
+      "policy/loss": 0.2885361313819885,
+      "step": 64
+    },
+    {
+      "completion_length": 246.125,
+      "epoch": 0.034666666666666665,
+      "grad_norm": 0.16669245064258575,
+      "learning_rate": 3e-06,
+      "loss": -0.0041,
+      "policy/loss": 0.20408771932125092,
+      "reward": 1.1510416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5885416567325592,
+      "rewards/correctness_reward_func_math": 0.5624999850988388,
+      "step": 65
+    },
+    {
+      "epoch": 0.0352,
+      "grad_norm": 0.21859301626682281,
+      "learning_rate": 3e-06,
+      "loss": -0.0013,
+      "policy/loss": 0.10099759697914124,
+      "step": 66
+    },
+    {
+      "epoch": 0.03573333333333333,
+      "grad_norm": 0.17133751511573792,
+      "learning_rate": 3e-06,
+      "loss": -0.0057,
+      "policy/loss": 0.20111273229122162,
+      "step": 67
+    },
+    {
+      "epoch": 0.03626666666666667,
+      "grad_norm": 0.23606789112091064,
+      "learning_rate": 3e-06,
+      "loss": -0.0011,
+      "policy/loss": 0.10094904899597168,
+      "step": 68
+    },
+    {
+      "completion_length": 246.2916717529297,
+      "epoch": 0.0368,
+      "grad_norm": 0.23300568759441376,
+      "learning_rate": 3e-06,
+      "loss": -0.0304,
+      "policy/loss": 0.04001743346452713,
+      "reward": 0.9062500298023224,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6354166567325592,
+      "rewards/correctness_reward_func_math": 0.2708333358168602,
+      "step": 69
+    },
+    {
+      "epoch": 0.037333333333333336,
+      "grad_norm": 0.21877847611904144,
+      "learning_rate": 3e-06,
+      "loss": -0.2528,
+      "policy/loss": -0.455715911462903,
+      "step": 70
+    },
+    {
+      "epoch": 0.037866666666666667,
+      "grad_norm": 0.21581768989562988,
+      "learning_rate": 3e-06,
+      "loss": -0.0308,
+      "policy/loss": 0.04030153062194586,
+      "step": 71
+    },
+    {
+      "epoch": 0.0384,
+      "grad_norm": 0.22192184627056122,
+      "learning_rate": 3e-06,
+      "loss": -0.2535,
+      "policy/loss": -0.45202592574059963,
+      "step": 72
+    },
+    {
+      "completion_length": 239.92708587646484,
+      "epoch": 0.038933333333333334,
+      "grad_norm": 0.21344998478889465,
+      "learning_rate": 3e-06,
+      "loss": 0.0544,
+      "policy/loss": -0.2639333406696096,
+      "reward": 1.1197916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.578125,
+      "rewards/correctness_reward_func_math": 0.5416666567325592,
+      "step": 73
+    },
+    {
+      "epoch": 0.039466666666666664,
+      "grad_norm": 0.16668258607387543,
+      "learning_rate": 3e-06,
+      "loss": 0.2541,
+      "policy/loss": 0.2643364667892456,
+      "step": 74
+    },
+    {
+      "epoch": 0.04,
+      "grad_norm": 0.19904448091983795,
+      "learning_rate": 3e-06,
+      "loss": 0.0544,
+      "policy/loss": -0.264262934528233,
+      "step": 75
+    },
+    {
+      "epoch": 0.04053333333333333,
+      "grad_norm": 0.17770980298519135,
+      "learning_rate": 3e-06,
+      "loss": 0.2538,
+      "policy/loss": 0.26365790888667107,
+      "step": 76
+    },
+    {
+      "completion_length": 241.21875,
+      "epoch": 0.04106666666666667,
+      "grad_norm": 0.16756409406661987,
+      "learning_rate": 3e-06,
+      "loss": -0.0046,
+      "policy/loss": 0.08975694328546524,
+      "reward": 1.1406250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.578125,
+      "rewards/correctness_reward_func_math": 0.5624999850988388,
+      "step": 77
+    },
+    {
+      "epoch": 0.0416,
+      "grad_norm": 0.26594436168670654,
+      "learning_rate": 3e-06,
+      "loss": 0.0667,
+      "policy/loss": -0.12227412313222885,
+      "step": 78
+    },
+    {
+      "epoch": 0.042133333333333335,
+      "grad_norm": 0.16517041623592377,
+      "learning_rate": 3e-06,
+      "loss": -0.0058,
+      "policy/loss": 0.08800341375172138,
+      "step": 79
+    },
+    {
+      "epoch": 0.042666666666666665,
+      "grad_norm": 0.225467249751091,
+      "learning_rate": 3e-06,
+      "loss": 0.066,
+      "policy/loss": -0.12077280506491661,
+      "step": 80
+    },
+    {
+      "completion_length": 241.11459350585938,
+      "epoch": 0.0432,
+      "grad_norm": 0.23417052626609802,
+      "learning_rate": 3e-06,
+      "loss": -0.0817,
+      "policy/loss": -0.03933607361449987,
+      "reward": 1.0416667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5833333134651184,
+      "rewards/correctness_reward_func_math": 0.4583333432674408,
+      "step": 81
+    },
+    {
+      "epoch": 0.04373333333333333,
+      "grad_norm": 0.21120339632034302,
+      "learning_rate": 3e-06,
+      "loss": -0.1249,
+      "policy/loss": -0.04809920491331354,
+      "step": 82
+    },
+    {
+      "epoch": 0.04426666666666667,
+      "grad_norm": 0.20807330310344696,
+      "learning_rate": 3e-06,
+      "loss": -0.0823,
+      "policy/loss": -0.03903044707394088,
+      "step": 83
+    },
+    {
+      "epoch": 0.0448,
+      "grad_norm": 0.20678380131721497,
+      "learning_rate": 3e-06,
+      "loss": -0.1251,
+      "policy/loss": -0.048280395618057526,
+      "step": 84
+    },
+    {
+      "completion_length": 244.2604217529297,
+      "epoch": 0.04533333333333334,
+      "grad_norm": 0.17263862490653992,
+      "learning_rate": 3e-06,
+      "loss": 0.0747,
+      "policy/loss": -0.03617490082979202,
+      "reward": 1.0312500298023224,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5729166567325592,
+      "rewards/correctness_reward_func_math": 0.4583333283662796,
+      "step": 85
+    },
+    {
+      "epoch": 0.04586666666666667,
+      "grad_norm": 0.22495436668395996,
+      "learning_rate": 3e-06,
+      "loss": -0.0467,
+      "policy/loss": -0.23247838020324707,
+      "step": 86
+    },
+    {
+      "epoch": 0.0464,
+      "grad_norm": 0.1758430004119873,
+      "learning_rate": 3e-06,
+      "loss": 0.0757,
+      "policy/loss": -0.030533596873283386,
+      "step": 87
+    },
+    {
+      "epoch": 0.046933333333333334,
+      "grad_norm": 0.23250040411949158,
+      "learning_rate": 3e-06,
+      "loss": -0.0463,
+      "policy/loss": -0.23382258415222168,
+      "step": 88
+    },
+    {
+      "completion_length": 252.06250762939453,
+      "epoch": 0.047466666666666664,
+      "grad_norm": 0.14873482286930084,
+      "learning_rate": 3e-06,
+      "loss": 0.0515,
+      "policy/loss": -0.021389774978160858,
+      "reward": 1.5364583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6197916567325592,
+      "rewards/correctness_reward_func_math": 0.9166666865348816,
+      "step": 89
+    },
+    {
+      "epoch": 0.048,
+      "grad_norm": 0.22680668532848358,
+      "learning_rate": 3e-06,
+      "loss": -0.0996,
+      "policy/loss": 0.2239882592111826,
+      "step": 90
+    },
+    {
+      "epoch": 0.04853333333333333,
+      "grad_norm": 0.14219488203525543,
+      "learning_rate": 3e-06,
+      "loss": 0.0506,
+      "policy/loss": -0.022728443145751953,
+      "step": 91
+    },
+    {
+      "epoch": 0.04906666666666667,
+      "grad_norm": 0.2535868287086487,
+      "learning_rate": 3e-06,
+      "loss": -0.1007,
+      "policy/loss": 0.223356818780303,
+      "step": 92
+    },
+    {
+      "completion_length": 231.67708587646484,
+      "epoch": 0.0496,
+      "grad_norm": 0.17225182056427002,
+      "learning_rate": 3e-06,
+      "loss": 0.087,
+      "policy/loss": -0.41347941756248474,
+      "reward": 1.5052083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6302083134651184,
+      "rewards/correctness_reward_func_math": 0.8750000298023224,
+      "step": 93
+    },
+    {
+      "epoch": 0.050133333333333335,
+      "grad_norm": 0.16623133420944214,
+      "learning_rate": 3e-06,
+      "loss": -0.0328,
+      "policy/loss": 0.10073413327336311,
+      "step": 94
+    },
+    {
+      "epoch": 0.050666666666666665,
+      "grad_norm": 0.1602189987897873,
+      "learning_rate": 3e-06,
+      "loss": 0.0861,
+      "policy/loss": -0.41538931988179684,
+      "step": 95
+    },
+    {
+      "epoch": 0.0512,
+      "grad_norm": 0.16254831850528717,
+      "learning_rate": 3e-06,
+      "loss": -0.0338,
+      "policy/loss": 0.09904193878173828,
+      "step": 96
+    },
+    {
+      "completion_length": 234.5729217529297,
+      "epoch": 0.05173333333333333,
+      "grad_norm": 0.16979831457138062,
+      "learning_rate": 3e-06,
+      "loss": 0.0319,
+      "policy/loss": -0.01932654157280922,
+      "reward": 1.1927083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.5416666716337204,
+      "step": 97
+    },
+    {
+      "epoch": 0.05226666666666667,
+      "grad_norm": 0.21723443269729614,
+      "learning_rate": 3e-06,
+      "loss": 0.0306,
+      "policy/loss": 0.13807255402207375,
+      "step": 98
+    },
+    {
+      "epoch": 0.0528,
+      "grad_norm": 0.16911055147647858,
+      "learning_rate": 3e-06,
+      "loss": 0.0315,
+      "policy/loss": -0.017486371099948883,
+      "step": 99
+    },
+    {
+      "epoch": 0.05333333333333334,
+      "grad_norm": 0.18764765560626984,
+      "learning_rate": 3e-06,
+      "loss": 0.0311,
+      "policy/loss": 0.13651283644139767,
+      "step": 100
+    },
+    {
+      "completion_length": 242.09375762939453,
+      "epoch": 0.05386666666666667,
+      "grad_norm": 0.2415752112865448,
+      "learning_rate": 3e-06,
+      "loss": -0.0679,
+      "policy/loss": 0.15059407953669535,
+      "reward": 1.3593750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.7083333283662796,
+      "step": 101
+    },
+    {
+      "epoch": 0.0544,
+      "grad_norm": 0.17402076721191406,
+      "learning_rate": 3e-06,
+      "loss": 0.0421,
+      "policy/loss": 0.12749002742382487,
+      "step": 102
+    },
+    {
+      "epoch": 0.054933333333333334,
+      "grad_norm": 0.21787986159324646,
+      "learning_rate": 3e-06,
+      "loss": -0.0688,
+      "policy/loss": 0.1485946565056082,
+      "step": 103
+    },
+    {
+      "epoch": 0.055466666666666664,
+      "grad_norm": 0.1745985448360443,
+      "learning_rate": 3e-06,
+      "loss": 0.0411,
+      "policy/loss": 0.1285392330537718,
+      "step": 104
+    },
+    {
+      "completion_length": 251.875,
+      "epoch": 0.056,
+      "grad_norm": 0.19410797953605652,
+      "learning_rate": 3e-06,
+      "loss": -0.0278,
+      "policy/loss": 8.574746068745753e-08,
+      "reward": 1.46875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6770833432674408,
+      "rewards/correctness_reward_func_math": 0.7916666865348816,
+      "step": 105
+    },
+    {
+      "epoch": 0.05653333333333333,
+      "grad_norm": 0.14488594233989716,
+      "learning_rate": 3e-06,
+      "loss": -0.1406,
+      "policy/loss": 9.973444825561728e-08,
+      "step": 106
+    },
+    {
+      "epoch": 0.05706666666666667,
+      "grad_norm": 0.18901671469211578,
+      "learning_rate": 3e-06,
+      "loss": -0.0275,
+      "policy/loss": 8.553086594531578e-08,
+      "step": 107
+    },
+    {
+      "epoch": 0.0576,
+      "grad_norm": 0.13076430559158325,
+      "learning_rate": 3e-06,
+      "loss": -0.1402,
+      "policy/loss": 9.951103230321223e-08,
+      "step": 108
+    },
+    {
+      "completion_length": 245.86458587646484,
+      "epoch": 0.058133333333333335,
+      "grad_norm": 0.18580268323421478,
+      "learning_rate": 3e-06,
+      "loss": 0.0087,
+      "policy/loss": -0.18717020004987717,
+      "reward": 1.3020833432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5729166865348816,
+      "rewards/correctness_reward_func_math": 0.7291666865348816,
+      "step": 109
+    },
+    {
+      "epoch": 0.058666666666666666,
+      "grad_norm": 0.210600808262825,
+      "learning_rate": 3e-06,
+      "loss": -0.0632,
+      "policy/loss": -0.4287007451057434,
+      "step": 110
+    },
+    {
+      "epoch": 0.0592,
+      "grad_norm": 0.2117600440979004,
+      "learning_rate": 3e-06,
+      "loss": 0.0065,
+      "policy/loss": -0.19087310880422592,
+      "step": 111
+    },
+    {
+      "epoch": 0.05973333333333333,
+      "grad_norm": 0.2067139893770218,
+      "learning_rate": 3e-06,
+      "loss": -0.065,
+      "policy/loss": -0.43042244762182236,
+      "step": 112
+    },
+    {
+      "completion_length": 246.14583587646484,
+      "epoch": 0.06026666666666667,
+      "grad_norm": 0.16639241576194763,
+      "learning_rate": 3e-06,
+      "loss": 0.0592,
+      "policy/loss": -0.161590675030034,
+      "reward": 1.2812500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6979166567325592,
+      "rewards/correctness_reward_func_math": 0.5833333432674408,
+      "step": 113
+    },
+    {
+      "epoch": 0.0608,
+      "grad_norm": 0.1712525486946106,
+      "learning_rate": 3e-06,
+      "loss": 0.0944,
+      "policy/loss": 0.1700624647597202,
+      "step": 114
+    },
+    {
+      "epoch": 0.06133333333333333,
+      "grad_norm": 0.17161071300506592,
+      "learning_rate": 3e-06,
+      "loss": 0.058,
+      "policy/loss": -0.16012598030746616,
+      "step": 115
+    },
+    {
+      "epoch": 0.06186666666666667,
+      "grad_norm": 0.16238321363925934,
+      "learning_rate": 3e-06,
+      "loss": 0.0934,
+      "policy/loss": 0.17073814297807033,
+      "step": 116
+    },
+    {
+      "completion_length": 251.4791717529297,
+      "epoch": 0.0624,
+      "grad_norm": 0.14962513744831085,
+      "learning_rate": 3e-06,
+      "loss": 0.0521,
+      "policy/loss": -0.44218730733582845,
+      "reward": 1.4166667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6458333134651184,
+      "rewards/correctness_reward_func_math": 0.7708333134651184,
+      "step": 117
+    },
+    {
+      "epoch": 0.06293333333333333,
+      "grad_norm": 0.15881341695785522,
+      "learning_rate": 3e-06,
+      "loss": 0.0072,
+      "policy/loss": -0.2324807717305788,
+      "step": 118
+    },
+    {
+      "epoch": 0.06346666666666667,
+      "grad_norm": 0.1459859311580658,
+      "learning_rate": 3e-06,
+      "loss": 0.0524,
+      "policy/loss": -0.44258439352104517,
+      "step": 119
+    },
+    {
+      "epoch": 0.064,
+      "grad_norm": 0.13922378420829773,
+      "learning_rate": 3e-06,
+      "loss": 0.0064,
+      "policy/loss": -0.23883009683111567,
+      "step": 120
+    },
+    {
+      "completion_length": 243.27083587646484,
+      "epoch": 0.06453333333333333,
+      "grad_norm": 0.23200690746307373,
+      "learning_rate": 3e-06,
+      "loss": -0.1004,
+      "policy/loss": 1.3466371129311483e-07,
+      "reward": 0.9531250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5989583432674408,
+      "rewards/correctness_reward_func_math": 0.3541666641831398,
+      "step": 121
+    },
+    {
+      "epoch": 0.06506666666666666,
+      "grad_norm": 0.19001558423042297,
+      "learning_rate": 3e-06,
+      "loss": -0.0017,
+      "policy/loss": 1.4731813280377537e-07,
+      "step": 122
+    },
+    {
+      "epoch": 0.0656,
+      "grad_norm": 0.44606658816337585,
+      "learning_rate": 3e-06,
+      "loss": -0.1017,
+      "policy/loss": 1.3465670889445391e-07,
+      "step": 123
+    },
+    {
+      "epoch": 0.06613333333333334,
+      "grad_norm": 0.17701329290866852,
+      "learning_rate": 3e-06,
+      "loss": -0.0021,
+      "policy/loss": 1.4726212782534276e-07,
+      "step": 124
+    },
+    {
+      "completion_length": 251.80209350585938,
+      "epoch": 0.06666666666666667,
+      "grad_norm": 0.1870919018983841,
+      "learning_rate": 3e-06,
+      "loss": -0.0511,
+      "policy/loss": 0.17864644527435303,
+      "reward": 1.484375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.671875,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 125
+    },
+    {
+      "epoch": 0.0672,
+      "grad_norm": 0.20933383703231812,
+      "learning_rate": 3e-06,
+      "loss": 0.1718,
+      "policy/loss": 0.11261439323425293,
+      "step": 126
+    },
+    {
+      "epoch": 0.06773333333333334,
+      "grad_norm": 0.18220266699790955,
+      "learning_rate": 3e-06,
+      "loss": -0.0518,
+      "policy/loss": 0.17746074497699738,
+      "step": 127
+    },
+    {
+      "epoch": 0.06826666666666667,
+      "grad_norm": 0.21195226907730103,
+      "learning_rate": 3e-06,
+      "loss": 0.1712,
+      "policy/loss": 0.1066715233027935,
+      "step": 128
+    },
+    {
+      "completion_length": 245.05209350585938,
+      "epoch": 0.0688,
+      "grad_norm": 0.14478085935115814,
+      "learning_rate": 3e-06,
+      "loss": 0.2445,
+      "policy/loss": -0.18012602627277374,
+      "reward": 1.1041666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.4166666716337204,
+      "step": 129
+    },
+    {
+      "epoch": 0.06933333333333333,
+      "grad_norm": 0.163246288895607,
+      "learning_rate": 3e-06,
+      "loss": 0.088,
+      "policy/loss": -0.03818912757560611,
+      "step": 130
+    },
+    {
+      "epoch": 0.06986666666666666,
+      "grad_norm": 0.13999663293361664,
+      "learning_rate": 3e-06,
+      "loss": 0.2433,
+      "policy/loss": -0.18185029923915863,
+      "step": 131
+    },
+    {
+      "epoch": 0.0704,
+      "grad_norm": 0.1706940233707428,
+      "learning_rate": 3e-06,
+      "loss": 0.088,
+      "policy/loss": -0.0393277476541698,
+      "step": 132
+    },
+    {
+      "completion_length": 238.43750762939453,
+      "epoch": 0.07093333333333333,
+      "grad_norm": 0.1640995889902115,
+      "learning_rate": 3e-06,
+      "loss": 0.0118,
+      "policy/loss": -0.188654427902744,
+      "reward": 1.359375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.609375,
+      "rewards/correctness_reward_func_math": 0.7500000149011612,
+      "step": 133
+    },
+    {
+      "epoch": 0.07146666666666666,
+      "grad_norm": 0.1892654448747635,
+      "learning_rate": 3e-06,
+      "loss": -0.0016,
+      "policy/loss": 0.20929016752674912,
+      "step": 134
+    },
+    {
+      "epoch": 0.072,
+      "grad_norm": 0.1590232104063034,
+      "learning_rate": 3e-06,
+      "loss": 0.0115,
+      "policy/loss": -0.18969442467339448,
+      "step": 135
+    },
+    {
+      "epoch": 0.07253333333333334,
+      "grad_norm": 0.18162208795547485,
+      "learning_rate": 3e-06,
+      "loss": -0.0018,
+      "policy/loss": 0.20950624925016115,
+      "step": 136
+    },
+    {
+      "completion_length": 240.43750762939453,
+      "epoch": 0.07306666666666667,
+      "grad_norm": 0.15719756484031677,
+      "learning_rate": 3e-06,
+      "loss": 0.0463,
+      "policy/loss": -0.17717686161048363,
+      "reward": 1.2968750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6302083432674408,
+      "rewards/correctness_reward_func_math": 0.6666666567325592,
+      "step": 137
+    },
+    {
+      "epoch": 0.0736,
+      "grad_norm": 0.15760871767997742,
+      "learning_rate": 3e-06,
+      "loss": -0.04,
+      "policy/loss": -0.5817977882790828,
+      "step": 138
+    },
+    {
+      "epoch": 0.07413333333333333,
+      "grad_norm": 0.15394710004329681,
+      "learning_rate": 3e-06,
+      "loss": 0.047,
+      "policy/loss": -0.1746198521073623,
+      "step": 139
+    },
+    {
+      "epoch": 0.07466666666666667,
+      "grad_norm": 0.16422191262245178,
+      "learning_rate": 3e-06,
+      "loss": -0.0403,
+      "policy/loss": -0.5815951920871854,
+      "step": 140
+    },
+    {
+      "completion_length": 241.08333587646484,
+      "epoch": 0.0752,
+      "grad_norm": 0.20093117654323578,
+      "learning_rate": 3e-06,
+      "loss": -0.2762,
+      "policy/loss": -0.5723820328712463,
+      "reward": 1.1666667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.5416666716337204,
+      "step": 141
+    },
+    {
+      "epoch": 0.07573333333333333,
+      "grad_norm": 0.17693877220153809,
+      "learning_rate": 3e-06,
+      "loss": -0.1741,
+      "policy/loss": -0.5187830477952957,
+      "step": 142
+    },
+    {
+      "epoch": 0.07626666666666666,
+      "grad_norm": 0.2006877213716507,
+      "learning_rate": 3e-06,
+      "loss": -0.2773,
+      "policy/loss": -0.5736844837665558,
+      "step": 143
+    },
+    {
+      "epoch": 0.0768,
+      "grad_norm": 0.1894705593585968,
+      "learning_rate": 3e-06,
+      "loss": -0.1755,
+      "policy/loss": -0.5210119336843491,
+      "step": 144
+    },
+    {
+      "completion_length": 245.67708587646484,
+      "epoch": 0.07733333333333334,
+      "grad_norm": 0.20710624754428864,
+      "learning_rate": 3e-06,
+      "loss": 0.0253,
+      "policy/loss": -0.08744821921566981,
+      "reward": 1.0677083432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416865348816,
+      "rewards/correctness_reward_func_math": 0.4166666716337204,
+      "step": 145
+    },
+    {
+      "epoch": 0.07786666666666667,
+      "grad_norm": 0.19527098536491394,
+      "learning_rate": 3e-06,
+      "loss": -0.0554,
+      "policy/loss": -0.15407992036477935,
+      "step": 146
+    },
+    {
+      "epoch": 0.0784,
+      "grad_norm": 0.21476508677005768,
+      "learning_rate": 3e-06,
+      "loss": 0.0251,
+      "policy/loss": -0.08439902446025016,
+      "step": 147
+    },
+    {
+      "epoch": 0.07893333333333333,
+      "grad_norm": 0.2004443258047104,
+      "learning_rate": 3e-06,
+      "loss": -0.0569,
+      "policy/loss": -0.15860990319781365,
+      "step": 148
+    },
+    {
+      "completion_length": 248.4479217529297,
+      "epoch": 0.07946666666666667,
+      "grad_norm": 0.1503647118806839,
+      "learning_rate": 3e-06,
+      "loss": 0.0126,
+      "policy/loss": -0.21599377803784137,
+      "reward": 1.3072917461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.640625,
+      "rewards/correctness_reward_func_math": 0.6666666865348816,
+      "step": 149
+    },
+    {
+      "epoch": 0.08,
+      "grad_norm": 0.1623055338859558,
+      "learning_rate": 3e-06,
+      "loss": 0.0932,
+      "policy/loss": -0.054830118797021044,
+      "step": 150
+    },
+    {
+      "epoch": 0.08053333333333333,
+      "grad_norm": 0.17495869100093842,
+      "learning_rate": 3e-06,
+      "loss": 0.0116,
+      "policy/loss": -0.2172851872049506,
+      "step": 151
+    },
+    {
+      "epoch": 0.08106666666666666,
+      "grad_norm": 0.39696913957595825,
+      "learning_rate": 3e-06,
+      "loss": 0.0929,
+      "policy/loss": -0.05583542547084619,
+      "step": 152
+    },
+    {
+      "completion_length": 238.77084350585938,
+      "epoch": 0.0816,
+      "grad_norm": 0.20325084030628204,
+      "learning_rate": 3e-06,
+      "loss": -0.165,
+      "policy/loss": -0.0112522062544258,
+      "reward": 1.4270833432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.7708333134651184,
+      "step": 153
+    },
+    {
+      "epoch": 0.08213333333333334,
+      "grad_norm": 0.19104507565498352,
+      "learning_rate": 3e-06,
+      "loss": 0.0203,
+      "policy/loss": 0.3923424800444266,
+      "step": 154
+    },
+    {
+      "epoch": 0.08266666666666667,
+      "grad_norm": 0.1926201432943344,
+      "learning_rate": 3e-06,
+      "loss": -0.1667,
+      "policy/loss": -0.018263615244677567,
+      "step": 155
+    },
+    {
+      "epoch": 0.0832,
+      "grad_norm": 0.2502761781215668,
+      "learning_rate": 3e-06,
+      "loss": 0.0195,
+      "policy/loss": 0.38940191465884055,
+      "step": 156
+    },
+    {
+      "completion_length": 243.50000762939453,
+      "epoch": 0.08373333333333334,
+      "grad_norm": 0.19198907911777496,
+      "learning_rate": 3e-06,
+      "loss": 0.0066,
+      "policy/loss": 0.13674150786778938,
+      "reward": 1.0781250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6614583432674408,
+      "rewards/correctness_reward_func_math": 0.4166666567325592,
+      "step": 157
+    },
+    {
+      "epoch": 0.08426666666666667,
+      "grad_norm": 0.24089176952838898,
+      "learning_rate": 3e-06,
+      "loss": -0.0621,
+      "policy/loss": -0.16569343647912405,
+      "step": 158
+    },
+    {
+      "epoch": 0.0848,
+      "grad_norm": 0.2007446438074112,
+      "learning_rate": 3e-06,
+      "loss": 0.0056,
+      "policy/loss": 0.13470405730192425,
+      "step": 159
+    },
+    {
+      "epoch": 0.08533333333333333,
+      "grad_norm": 0.2479970008134842,
+      "learning_rate": 3e-06,
+      "loss": -0.0637,
+      "policy/loss": -0.17106989464721778,
+      "step": 160
+    },
+    {
+      "completion_length": 247.56250762939453,
+      "epoch": 0.08586666666666666,
+      "grad_norm": 0.23994049429893494,
+      "learning_rate": 3e-06,
+      "loss": -0.1204,
+      "policy/loss": -0.1656414233148098,
+      "reward": 1.5312500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8125000298023224,
+      "step": 161
+    },
+    {
+      "epoch": 0.0864,
+      "grad_norm": 0.24293623864650726,
+      "learning_rate": 3e-06,
+      "loss": 0.0538,
+      "policy/loss": 0.42442003823816776,
+      "step": 162
+    },
+    {
+      "epoch": 0.08693333333333333,
+      "grad_norm": 0.25687435269355774,
+      "learning_rate": 3e-06,
+      "loss": -0.1196,
+      "policy/loss": -0.15939903259277344,
+      "step": 163
+    },
+    {
+      "epoch": 0.08746666666666666,
+      "grad_norm": 0.2583344876766205,
+      "learning_rate": 3e-06,
+      "loss": 0.0531,
+      "policy/loss": 0.4205309543758631,
+      "step": 164
+    },
+    {
+      "completion_length": 246.95833587646484,
+      "epoch": 0.088,
+      "grad_norm": 0.19497521221637726,
+      "learning_rate": 3e-06,
+      "loss": -0.0558,
+      "policy/loss": -0.3753738012164831,
+      "reward": 1.3802083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.6666666567325592,
+      "step": 165
+    },
+    {
+      "epoch": 0.08853333333333334,
+      "grad_norm": 0.20267871022224426,
+      "learning_rate": 3e-06,
+      "loss": 0.1557,
+      "policy/loss": -0.1712660826742649,
+      "step": 166
+    },
+    {
+      "epoch": 0.08906666666666667,
+      "grad_norm": 0.22812624275684357,
+      "learning_rate": 3e-06,
+      "loss": -0.0565,
+      "policy/loss": -0.38266682624816895,
+      "step": 167
+    },
+    {
+      "epoch": 0.0896,
+      "grad_norm": 0.2134062796831131,
+      "learning_rate": 3e-06,
+      "loss": 0.1552,
+      "policy/loss": -0.17667826265096664,
+      "step": 168
+    },
+    {
+      "completion_length": 239.6041717529297,
+      "epoch": 0.09013333333333333,
+      "grad_norm": 0.20163583755493164,
+      "learning_rate": 3e-06,
+      "loss": 0.0109,
+      "policy/loss": 0.14976216852664948,
+      "reward": 1.2447916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6614583432674408,
+      "rewards/correctness_reward_func_math": 0.5833333432674408,
+      "step": 169
+    },
+    {
+      "epoch": 0.09066666666666667,
+      "grad_norm": 0.1710713654756546,
+      "learning_rate": 3e-06,
+      "loss": 0.1258,
+      "policy/loss": 0.686280369758606,
+      "step": 170
+    },
+    {
+      "epoch": 0.0912,
+      "grad_norm": 0.17095541954040527,
+      "learning_rate": 3e-06,
+      "loss": 0.0101,
+      "policy/loss": 0.14622743427753448,
+      "step": 171
+    },
+    {
+      "epoch": 0.09173333333333333,
+      "grad_norm": 0.16383086144924164,
+      "learning_rate": 3e-06,
+      "loss": 0.1261,
+      "policy/loss": 0.6873235702514648,
+      "step": 172
+    },
+    {
+      "completion_length": 239.92709350585938,
+      "epoch": 0.09226666666666666,
+      "grad_norm": 0.15347491204738617,
+      "learning_rate": 3e-06,
+      "loss": 0.0937,
+      "policy/loss": 0.4111483171582222,
+      "reward": 1.6250000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333432674408,
+      "rewards/correctness_reward_func_math": 0.9166666567325592,
+      "step": 173
+    },
+    {
+      "epoch": 0.0928,
+      "grad_norm": 0.16088056564331055,
+      "learning_rate": 3e-06,
+      "loss": 0.1638,
+      "policy/loss": 0.41885845363140106,
+      "step": 174
+    },
+    {
+      "epoch": 0.09333333333333334,
+      "grad_norm": 0.15280427038669586,
+      "learning_rate": 3e-06,
+      "loss": 0.0929,
+      "policy/loss": 0.40875808894634247,
+      "step": 175
+    },
+    {
+      "epoch": 0.09386666666666667,
+      "grad_norm": 0.15237613022327423,
+      "learning_rate": 3e-06,
+      "loss": 0.1624,
+      "policy/loss": 0.41541149467229843,
+      "step": 176
+    },
+    {
+      "completion_length": 245.3854217529297,
+      "epoch": 0.0944,
+      "grad_norm": 0.21077562868595123,
+      "learning_rate": 3e-06,
+      "loss": -0.0679,
+      "policy/loss": -0.14120862850412408,
+      "reward": 0.838541716337204,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5260416567325592,
+      "rewards/correctness_reward_func_math": 0.3125000074505806,
+      "step": 177
+    },
+    {
+      "epoch": 0.09493333333333333,
+      "grad_norm": 0.20059771835803986,
+      "learning_rate": 3e-06,
+      "loss": 0.1574,
+      "policy/loss": 0.39109204509001927,
+      "step": 178
+    },
+    {
+      "epoch": 0.09546666666666667,
+      "grad_norm": 0.22224459052085876,
+      "learning_rate": 3e-06,
+      "loss": -0.0689,
+      "policy/loss": -0.14299199939181761,
+      "step": 179
+    },
+    {
+      "epoch": 0.096,
+      "grad_norm": 0.21497687697410583,
+      "learning_rate": 3e-06,
+      "loss": 0.1574,
+      "policy/loss": 0.3902242907287743,
+      "step": 180
+    },
+    {
+      "completion_length": 220.89583587646484,
+      "epoch": 0.09653333333333333,
+      "grad_norm": 0.16718493402004242,
+      "learning_rate": 3e-06,
+      "loss": 0.027,
+      "policy/loss": 0.3067523835698971,
+      "reward": 1.3281250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6614583432674408,
+      "rewards/correctness_reward_func_math": 0.6666666865348816,
+      "step": 181
+    },
+    {
+      "epoch": 0.09706666666666666,
+      "grad_norm": 0.15653477609157562,
+      "learning_rate": 3e-06,
+      "loss": -0.0773,
+      "policy/loss": 0.28450418196725025,
+      "step": 182
+    },
+    {
+      "epoch": 0.0976,
+      "grad_norm": 0.16930420696735382,
+      "learning_rate": 3e-06,
+      "loss": 0.028,
+      "policy/loss": 0.30576392995538626,
+      "step": 183
+    },
+    {
+      "epoch": 0.09813333333333334,
+      "grad_norm": 0.14894235134124756,
+      "learning_rate": 3e-06,
+      "loss": -0.0775,
+      "policy/loss": 0.28511709651814243,
+      "step": 184
+    },
+    {
+      "completion_length": 228.27083587646484,
+      "epoch": 0.09866666666666667,
+      "grad_norm": 0.20474110543727875,
+      "learning_rate": 3e-06,
+      "loss": 0.0336,
+      "policy/loss": 1.3163407075467148e-07,
+      "reward": 1.4322916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 185
+    },
+    {
+      "epoch": 0.0992,
+      "grad_norm": 0.18689489364624023,
+      "learning_rate": 3e-06,
+      "loss": 0.0165,
+      "policy/loss": 1.5078100545906636e-07,
+      "step": 186
+    },
+    {
+      "epoch": 0.09973333333333333,
+      "grad_norm": 0.19402816891670227,
+      "learning_rate": 3e-06,
+      "loss": 0.0327,
+      "policy/loss": 1.3157255907003673e-07,
+      "step": 187
+    },
+    {
+      "epoch": 0.10026666666666667,
+      "grad_norm": 0.17962224781513214,
+      "learning_rate": 3e-06,
+      "loss": 0.0146,
+      "policy/loss": 1.5065154457261087e-07,
+      "step": 188
+    },
+    {
+      "completion_length": 235.0104217529297,
+      "epoch": 0.1008,
+      "grad_norm": 0.17147351801395416,
+      "learning_rate": 3e-06,
+      "loss": 0.1876,
+      "policy/loss": 0.6175396889448166,
+      "reward": 1.0312500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.59375,
+      "rewards/correctness_reward_func_math": 0.4375,
+      "step": 189
+    },
+    {
+      "epoch": 0.10133333333333333,
+      "grad_norm": 0.19576120376586914,
+      "learning_rate": 3e-06,
+      "loss": 0.1786,
+      "policy/loss": 0.4932377729564905,
+      "step": 190
+    },
+    {
+      "epoch": 0.10186666666666666,
+      "grad_norm": 0.16359369456768036,
+      "learning_rate": 3e-06,
+      "loss": 0.187,
+      "policy/loss": 0.6148133277893066,
+      "step": 191
+    },
+    {
+      "epoch": 0.1024,
+      "grad_norm": 0.1803278774023056,
+      "learning_rate": 3e-06,
+      "loss": 0.1771,
+      "policy/loss": 0.4838356110267341,
+      "step": 192
+    },
+    {
+      "completion_length": 245.5104217529297,
+      "epoch": 0.10293333333333334,
+      "grad_norm": 0.16759775578975677,
+      "learning_rate": 3e-06,
+      "loss": 0.055,
+      "policy/loss": 0.33043980563459385,
+      "reward": 1.1614583432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.703125,
+      "rewards/correctness_reward_func_math": 0.4583333432674408,
+      "step": 193
+    },
+    {
+      "epoch": 0.10346666666666667,
+      "grad_norm": 0.19072411954402924,
+      "learning_rate": 3e-06,
+      "loss": 0.0743,
+      "policy/loss": -0.11677173565816545,
+      "step": 194
+    },
+    {
+      "epoch": 0.104,
+      "grad_norm": 0.20836253464221954,
+      "learning_rate": 3e-06,
+      "loss": 0.0539,
+      "policy/loss": 0.3281818922399289,
+      "step": 195
+    },
+    {
+      "epoch": 0.10453333333333334,
+      "grad_norm": 0.19410809874534607,
+      "learning_rate": 3e-06,
+      "loss": 0.0724,
+      "policy/loss": -0.11836121280921574,
+      "step": 196
+    },
+    {
+      "completion_length": 245.0729217529297,
+      "epoch": 0.10506666666666667,
+      "grad_norm": 0.21460328996181488,
+      "learning_rate": 3e-06,
+      "loss": -0.1176,
+      "policy/loss": 0.07936147600412369,
+      "reward": 1.1041666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6041666716337204,
+      "rewards/correctness_reward_func_math": 0.4999999850988388,
+      "step": 197
+    },
+    {
+      "epoch": 0.1056,
+      "grad_norm": 0.2519777715206146,
+      "learning_rate": 3e-06,
+      "loss": -0.0856,
+      "policy/loss": -0.12758401408791542,
+      "step": 198
+    },
+    {
+      "epoch": 0.10613333333333333,
+      "grad_norm": 0.1798715591430664,
+      "learning_rate": 3e-06,
+      "loss": -0.118,
+      "policy/loss": 0.07840218767523766,
+      "step": 199
+    },
+    {
+      "epoch": 0.10666666666666667,
+      "grad_norm": 0.21773001551628113,
+      "learning_rate": 3e-06,
+      "loss": -0.0871,
+      "policy/loss": -0.13118240237236023,
+      "step": 200
+    },
+    {
+      "completion_length": 242.48958587646484,
+      "epoch": 0.1072,
+      "grad_norm": 0.1787147969007492,
+      "learning_rate": 3e-06,
+      "loss": 0.2088,
+      "policy/loss": 0.13361368793994188,
+      "reward": 1.3125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333134651184,
+      "rewards/correctness_reward_func_math": 0.6041666716337204,
+      "step": 201
+    },
+    {
+      "epoch": 0.10773333333333333,
+      "grad_norm": 0.15103591978549957,
+      "learning_rate": 3e-06,
+      "loss": 0.1179,
+      "policy/loss": -0.15140119564603083,
+      "step": 202
+    },
+    {
+      "epoch": 0.10826666666666666,
+      "grad_norm": 0.2115238457918167,
+      "learning_rate": 3e-06,
+      "loss": 0.2078,
+      "policy/loss": 0.1309562250971794,
+      "step": 203
+    },
+    {
+      "epoch": 0.1088,
+      "grad_norm": 0.1507405787706375,
+      "learning_rate": 3e-06,
+      "loss": 0.1171,
+      "policy/loss": -0.15156680636573583,
+      "step": 204
+    },
+    {
+      "completion_length": 244.375,
+      "epoch": 0.10933333333333334,
+      "grad_norm": 0.19619819521903992,
+      "learning_rate": 3e-06,
+      "loss": 0.0628,
+      "policy/loss": 0.15721260523613978,
+      "reward": 1.1822916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6197916567325592,
+      "rewards/correctness_reward_func_math": 0.5625000149011612,
+      "step": 205
+    },
+    {
+      "epoch": 0.10986666666666667,
+      "grad_norm": 0.17184896767139435,
+      "learning_rate": 3e-06,
+      "loss": -0.011,
+      "policy/loss": 0.4386347357515703,
+      "step": 206
+    },
+    {
+      "epoch": 0.1104,
+      "grad_norm": 0.2206937074661255,
+      "learning_rate": 3e-06,
+      "loss": 0.0618,
+      "policy/loss": 0.15651119279709036,
+      "step": 207
+    },
+    {
+      "epoch": 0.11093333333333333,
+      "grad_norm": 0.17822307348251343,
+      "learning_rate": 3e-06,
+      "loss": -0.0122,
+      "policy/loss": 0.43496079485504424,
+      "step": 208
+    },
+    {
+      "completion_length": 246.1666717529297,
+      "epoch": 0.11146666666666667,
+      "grad_norm": 0.22039896249771118,
+      "learning_rate": 3e-06,
+      "loss": -0.0367,
+      "policy/loss": 1.4303645912150387e-07,
+      "reward": 1.28125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6770833432674408,
+      "rewards/correctness_reward_func_math": 0.6041666641831398,
+      "step": 209
+    },
+    {
+      "epoch": 0.112,
+      "grad_norm": 0.36158695816993713,
+      "learning_rate": 3e-06,
+      "loss": 0.0591,
+      "policy/loss": 1.5799814434558357e-07,
+      "step": 210
+    },
+    {
+      "epoch": 0.11253333333333333,
+      "grad_norm": 0.22058650851249695,
+      "learning_rate": 3e-06,
+      "loss": -0.0387,
+      "policy/loss": 1.429513183381914e-07,
+      "step": 211
+    },
+    {
+      "epoch": 0.11306666666666666,
+      "grad_norm": 0.4398540258407593,
+      "learning_rate": 3e-06,
+      "loss": 0.0578,
+      "policy/loss": 1.578616242170483e-07,
+      "step": 212
+    },
+    {
+      "completion_length": 245.80209350585938,
+      "epoch": 0.1136,
+      "grad_norm": 0.1663266271352768,
+      "learning_rate": 3e-06,
+      "loss": -0.0411,
+      "policy/loss": -0.40684649101166315,
+      "reward": 1.6770833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6770833134651184,
+      "rewards/correctness_reward_func_math": 0.9999999701976776,
+      "step": 213
+    },
+    {
+      "epoch": 0.11413333333333334,
+      "grad_norm": 0.17266380786895752,
+      "learning_rate": 3e-06,
+      "loss": 0.0166,
+      "policy/loss": -0.0678827189163087,
+      "step": 214
+    },
+    {
+      "epoch": 0.11466666666666667,
+      "grad_norm": 0.16179528832435608,
+      "learning_rate": 3e-06,
+      "loss": -0.0412,
+      "policy/loss": -0.4069986019977847,
+      "step": 215
+    },
+    {
+      "epoch": 0.1152,
+      "grad_norm": 0.17449289560317993,
+      "learning_rate": 3e-06,
+      "loss": 0.015,
+      "policy/loss": -0.0708675556850551,
+      "step": 216
+    },
+    {
+      "completion_length": 249.2604217529297,
+      "epoch": 0.11573333333333333,
+      "grad_norm": 0.16052700579166412,
+      "learning_rate": 3e-06,
+      "loss": -0.1301,
+      "policy/loss": 8.387050343117153e-08,
+      "reward": 1.2343750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.5208333283662796,
+      "step": 217
+    },
+    {
+      "epoch": 0.11626666666666667,
+      "grad_norm": 0.17012521624565125,
+      "learning_rate": 3e-06,
+      "loss": 0.0065,
+      "policy/loss": 1.1005875322211978e-07,
+      "step": 218
+    },
+    {
+      "epoch": 0.1168,
+      "grad_norm": 0.16360628604888916,
+      "learning_rate": 3e-06,
+      "loss": -0.1311,
+      "policy/loss": 8.403234019738193e-08,
+      "step": 219
+    },
+    {
+      "epoch": 0.11733333333333333,
+      "grad_norm": 0.16951008141040802,
+      "learning_rate": 3e-06,
+      "loss": 0.0061,
+      "policy/loss": 1.101523459112741e-07,
+      "step": 220
+    },
+    {
+      "completion_length": 244.21875762939453,
+      "epoch": 0.11786666666666666,
+      "grad_norm": 0.1686982959508896,
+      "learning_rate": 3e-06,
+      "loss": 0.1098,
+      "policy/loss": 0.47729112364411286,
+      "reward": 1.010416716337204,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6354166865348816,
+      "rewards/correctness_reward_func_math": 0.375,
+      "step": 221
+    },
+    {
+      "epoch": 0.1184,
+      "grad_norm": 0.15598666667938232,
+      "learning_rate": 3e-06,
+      "loss": -0.0337,
+      "policy/loss": -0.39744446935991284,
+      "step": 222
+    },
+    {
+      "epoch": 0.11893333333333334,
+      "grad_norm": 0.17785470187664032,
+      "learning_rate": 3e-06,
+      "loss": 0.1086,
+      "policy/loss": 0.4774944947145965,
+      "step": 223
+    },
+    {
+      "epoch": 0.11946666666666667,
+      "grad_norm": 0.16505518555641174,
+      "learning_rate": 3e-06,
+      "loss": -0.0352,
+      "policy/loss": -0.3997157937995368,
+      "step": 224
+    },
+    {
+      "completion_length": 226.18750762939453,
+      "epoch": 0.12,
+      "grad_norm": 0.19105078279972076,
+      "learning_rate": 3e-06,
+      "loss": 0.0295,
+      "policy/loss": -0.12086368352174759,
+      "reward": 1.4583333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333134651184,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 225
+    },
+    {
+      "epoch": 0.12053333333333334,
+      "grad_norm": 0.21446490287780762,
+      "learning_rate": 3e-06,
+      "loss": 0.0332,
+      "policy/loss": 0.07203349471092224,
+      "step": 226
+    },
+    {
+      "epoch": 0.12106666666666667,
+      "grad_norm": 0.1930609792470932,
+      "learning_rate": 3e-06,
+      "loss": 0.03,
+      "policy/loss": -0.12034983560442924,
+      "step": 227
+    },
+    {
+      "epoch": 0.1216,
+      "grad_norm": 0.1865423172712326,
+      "learning_rate": 3e-06,
+      "loss": 0.0326,
+      "policy/loss": 0.07144516333937645,
+      "step": 228
+    },
+    {
+      "completion_length": 237.31250762939453,
+      "epoch": 0.12213333333333333,
+      "grad_norm": 0.23347723484039307,
+      "learning_rate": 3e-06,
+      "loss": -0.0806,
+      "policy/loss": 6.558667564604548e-08,
+      "reward": 1.234375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.5208333358168602,
+      "step": 229
+    },
+    {
+      "epoch": 0.12266666666666666,
+      "grad_norm": 0.16616100072860718,
+      "learning_rate": 3e-06,
+      "loss": -0.0755,
+      "policy/loss": 5.158822702355792e-08,
+      "step": 230
+    },
+    {
+      "epoch": 0.1232,
+      "grad_norm": 0.2065654844045639,
+      "learning_rate": 3e-06,
+      "loss": -0.0808,
+      "policy/loss": 6.563587362506951e-08,
+      "step": 231
+    },
+    {
+      "epoch": 0.12373333333333333,
+      "grad_norm": 0.182975634932518,
+      "learning_rate": 3e-06,
+      "loss": -0.0753,
+      "policy/loss": 5.1566736658514856e-08,
+      "step": 232
+    },
+    {
+      "completion_length": 238.1666717529297,
+      "epoch": 0.12426666666666666,
+      "grad_norm": 0.19350838661193848,
+      "learning_rate": 3e-06,
+      "loss": 0.1232,
+      "policy/loss": 0.1293882798090067,
+      "reward": 1.265625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.703125,
+      "rewards/correctness_reward_func_math": 0.5625,
+      "step": 233
+    },
+    {
+      "epoch": 0.1248,
+      "grad_norm": 0.2704569697380066,
+      "learning_rate": 3e-06,
+      "loss": 0.1449,
+      "policy/loss": -0.15168937771689883,
+      "step": 234
+    },
+    {
+      "epoch": 0.12533333333333332,
+      "grad_norm": 0.18950939178466797,
+      "learning_rate": 3e-06,
+      "loss": 0.1228,
+      "policy/loss": 0.126039228917314,
+      "step": 235
+    },
+    {
+      "epoch": 0.12586666666666665,
+      "grad_norm": 0.18691076338291168,
+      "learning_rate": 3e-06,
+      "loss": 0.144,
+      "policy/loss": -0.15267191555555115,
+      "step": 236
+    },
+    {
+      "completion_length": 246.34375,
+      "epoch": 0.1264,
+      "grad_norm": 0.19988317787647247,
+      "learning_rate": 3e-06,
+      "loss": -0.037,
+      "policy/loss": -0.22487123031169176,
+      "reward": 1.2760417461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.5624999850988388,
+      "step": 237
+    },
+    {
+      "epoch": 0.12693333333333334,
+      "grad_norm": 0.31931889057159424,
+      "learning_rate": 3e-06,
+      "loss": 0.0237,
+      "policy/loss": 0.4042653050273657,
+      "step": 238
+    },
+    {
+      "epoch": 0.12746666666666667,
+      "grad_norm": 0.18974639475345612,
+      "learning_rate": 3e-06,
+      "loss": -0.0375,
+      "policy/loss": -0.22422014735639095,
+      "step": 239
+    },
+    {
+      "epoch": 0.128,
+      "grad_norm": 0.2391069084405899,
+      "learning_rate": 3e-06,
+      "loss": 0.0225,
+      "policy/loss": 0.40320491977036,
+      "step": 240
+    },
+    {
+      "completion_length": 245.7291717529297,
+      "epoch": 0.12853333333333333,
+      "grad_norm": 0.22687625885009766,
+      "learning_rate": 3e-06,
+      "loss": -0.0229,
+      "policy/loss": 0.07021526142588641,
+      "reward": 1.421875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.7083333432674408,
+      "step": 241
+    },
+    {
+      "epoch": 0.12906666666666666,
+      "grad_norm": 0.19379393756389618,
+      "learning_rate": 3e-06,
+      "loss": -0.072,
+      "policy/loss": -0.164413606964807,
+      "step": 242
+    },
+    {
+      "epoch": 0.1296,
+      "grad_norm": 0.2759920656681061,
+      "learning_rate": 3e-06,
+      "loss": -0.0243,
+      "policy/loss": 0.06729297978574778,
+      "step": 243
+    },
+    {
+      "epoch": 0.13013333333333332,
+      "grad_norm": 0.18662793934345245,
+      "learning_rate": 3e-06,
+      "loss": -0.0726,
+      "policy/loss": -0.16485341474380277,
+      "step": 244
+    },
+    {
+      "completion_length": 228.8541717529297,
+      "epoch": 0.13066666666666665,
+      "grad_norm": 0.169996976852417,
+      "learning_rate": 3e-06,
+      "loss": -0.1551,
+      "policy/loss": -0.289961613714695,
+      "reward": 1.3541666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333432674408,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 245
+    },
+    {
+      "epoch": 0.1312,
+      "grad_norm": 0.19478197395801544,
+      "learning_rate": 3e-06,
+      "loss": -0.0917,
+      "policy/loss": 0.303764671087265,
+      "step": 246
+    },
+    {
+      "epoch": 0.13173333333333334,
+      "grad_norm": 0.16889190673828125,
+      "learning_rate": 3e-06,
+      "loss": -0.1548,
+      "policy/loss": -0.2881269231438637,
+      "step": 247
+    },
+    {
+      "epoch": 0.13226666666666667,
+      "grad_norm": 0.1830751597881317,
+      "learning_rate": 3e-06,
+      "loss": -0.0923,
+      "policy/loss": 0.30230512097477913,
+      "step": 248
+    },
+    {
+      "completion_length": 242.1979217529297,
+      "epoch": 0.1328,
+      "grad_norm": 0.17263826727867126,
+      "learning_rate": 3e-06,
+      "loss": 0.1184,
+      "policy/loss": 0.012345903511683076,
+      "reward": 0.9739583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5572916567325592,
+      "rewards/correctness_reward_func_math": 0.4166666716337204,
+      "step": 249
+    },
+    {
+      "epoch": 0.13333333333333333,
+      "grad_norm": 0.1888904720544815,
+      "learning_rate": 3e-06,
+      "loss": 0.0079,
+      "policy/loss": -0.12822893634165666,
+      "step": 250
+    },
+    {
+      "epoch": 0.13386666666666666,
+      "grad_norm": 0.17100490629673004,
+      "learning_rate": 3e-06,
+      "loss": 0.1172,
+      "policy/loss": 0.013934924236895796,
+      "step": 251
+    },
+    {
+      "epoch": 0.1344,
+      "grad_norm": 0.2251415252685547,
+      "learning_rate": 3e-06,
+      "loss": 0.0068,
+      "policy/loss": -0.1285824366145647,
+      "step": 252
+    },
+    {
+      "completion_length": 240.23959350585938,
+      "epoch": 0.13493333333333332,
+      "grad_norm": 0.2815396785736084,
+      "learning_rate": 3e-06,
+      "loss": 0.0133,
+      "policy/loss": -0.43287701152296876,
+      "reward": 1.1927083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.5416666716337204,
+      "step": 253
+    },
+    {
+      "epoch": 0.13546666666666668,
+      "grad_norm": 0.21855133771896362,
+      "learning_rate": 3e-06,
+      "loss": 0.0559,
+      "policy/loss": 0.2768915218762231,
+      "step": 254
+    },
+    {
+      "epoch": 0.136,
+      "grad_norm": 0.5586273670196533,
+      "learning_rate": 3e-06,
+      "loss": 0.0115,
+      "policy/loss": -0.4359213187862636,
+      "step": 255
+    },
+    {
+      "epoch": 0.13653333333333334,
+      "grad_norm": 0.2116006761789322,
+      "learning_rate": 3e-06,
+      "loss": 0.0543,
+      "policy/loss": 0.27313249534542194,
+      "step": 256
+    },
+    {
+      "completion_length": 243.52084350585938,
+      "epoch": 0.13706666666666667,
+      "grad_norm": 0.2149346023797989,
+      "learning_rate": 3e-06,
+      "loss": -0.0666,
+      "policy/loss": -0.5429651439189911,
+      "reward": 1.1510416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.671875,
+      "rewards/correctness_reward_func_math": 0.4791666716337204,
+      "step": 257
+    },
+    {
+      "epoch": 0.1376,
+      "grad_norm": 0.28507742285728455,
+      "learning_rate": 3e-06,
+      "loss": -0.0573,
+      "policy/loss": -0.5261799097061157,
+      "step": 258
+    },
+    {
+      "epoch": 0.13813333333333333,
+      "grad_norm": 0.20848101377487183,
+      "learning_rate": 3e-06,
+      "loss": -0.067,
+      "policy/loss": -0.544669583439827,
+      "step": 259
+    },
+    {
+      "epoch": 0.13866666666666666,
+      "grad_norm": 0.2138155996799469,
+      "learning_rate": 3e-06,
+      "loss": -0.0598,
+      "policy/loss": -0.5282830521464348,
+      "step": 260
+    },
+    {
+      "completion_length": 236.53125762939453,
+      "epoch": 0.1392,
+      "grad_norm": 0.22861425578594208,
+      "learning_rate": 3e-06,
+      "loss": -0.015,
+      "policy/loss": 0.10509785264730453,
+      "reward": 1.072916716337204,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.4166666716337204,
+      "step": 261
+    },
+    {
+      "epoch": 0.13973333333333332,
+      "grad_norm": 0.163916677236557,
+      "learning_rate": 3e-06,
+      "loss": -0.163,
+      "policy/loss": 0.12160508707165718,
+      "step": 262
+    },
+    {
+      "epoch": 0.14026666666666668,
+      "grad_norm": 0.22389155626296997,
+      "learning_rate": 3e-06,
+      "loss": -0.0161,
+      "policy/loss": 0.10276631824672222,
+      "step": 263
+    },
+    {
+      "epoch": 0.1408,
+      "grad_norm": 0.17089462280273438,
+      "learning_rate": 3e-06,
+      "loss": -0.1629,
+      "policy/loss": 0.12141638621687889,
+      "step": 264
+    },
+    {
+      "completion_length": 232.15625762939453,
+      "epoch": 0.14133333333333334,
+      "grad_norm": 0.18738433718681335,
+      "learning_rate": 3e-06,
+      "loss": -0.0298,
+      "policy/loss": 0.13852850867318267,
+      "reward": 1.2552083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.6041666716337204,
+      "step": 265
+    },
+    {
+      "epoch": 0.14186666666666667,
+      "grad_norm": 0.23036138713359833,
+      "learning_rate": 3e-06,
+      "loss": 0.0667,
+      "policy/loss": -0.23110564609249096,
+      "step": 266
+    },
+    {
+      "epoch": 0.1424,
+      "grad_norm": 0.19782599806785583,
+      "learning_rate": 3e-06,
+      "loss": -0.0306,
+      "policy/loss": 0.1380127050773865,
+      "step": 267
+    },
+    {
+      "epoch": 0.14293333333333333,
+      "grad_norm": 0.22464993596076965,
+      "learning_rate": 3e-06,
+      "loss": 0.0658,
+      "policy/loss": -0.22946252496262076,
+      "step": 268
+    },
+    {
+      "completion_length": 236.52083587646484,
+      "epoch": 0.14346666666666666,
+      "grad_norm": 0.21544314920902252,
+      "learning_rate": 3e-06,
+      "loss": -0.0435,
+      "policy/loss": 0.21933392329654566,
+      "reward": 1.5208333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6666666865348816,
+      "rewards/correctness_reward_func_math": 0.8541666865348816,
+      "step": 269
+    },
+    {
+      "epoch": 0.144,
+      "grad_norm": 0.22398751974105835,
+      "learning_rate": 3e-06,
+      "loss": -0.0733,
+      "policy/loss": 0.13499312679543607,
+      "step": 270
+    },
+    {
+      "epoch": 0.14453333333333335,
+      "grad_norm": 0.2145986706018448,
+      "learning_rate": 3e-06,
+      "loss": -0.0458,
+      "policy/loss": 0.22035375877531038,
+      "step": 271
+    },
+    {
+      "epoch": 0.14506666666666668,
+      "grad_norm": 0.20699889957904816,
+      "learning_rate": 3e-06,
+      "loss": -0.0745,
+      "policy/loss": 0.13548379237445118,
+      "step": 272
+    },
+    {
+      "completion_length": 225.7291717529297,
+      "epoch": 0.1456,
+      "grad_norm": 0.1854046732187271,
+      "learning_rate": 3e-06,
+      "loss": -0.0644,
+      "policy/loss": -0.0865171949079695,
+      "reward": 1.3958333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6666666567325592,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 273
+    },
+    {
+      "epoch": 0.14613333333333334,
+      "grad_norm": 0.17359668016433716,
+      "learning_rate": 3e-06,
+      "loss": -0.1002,
+      "policy/loss": 0.06539462845690025,
+      "step": 274
+    },
+    {
+      "epoch": 0.14666666666666667,
+      "grad_norm": 0.1708374172449112,
+      "learning_rate": 3e-06,
+      "loss": -0.0655,
+      "policy/loss": -0.09002277501354428,
+      "step": 275
+    },
+    {
+      "epoch": 0.1472,
+      "grad_norm": 0.18366111814975739,
+      "learning_rate": 3e-06,
+      "loss": -0.1004,
+      "policy/loss": 0.06363140382835297,
+      "step": 276
+    },
+    {
+      "completion_length": 235.64583587646484,
+      "epoch": 0.14773333333333333,
+      "grad_norm": 0.19279073178768158,
+      "learning_rate": 3e-06,
+      "loss": -0.0249,
+      "policy/loss": 1.2076397837290642e-07,
+      "reward": 1.3229166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.53125,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 277
+    },
+    {
+      "epoch": 0.14826666666666666,
+      "grad_norm": 0.21259015798568726,
+      "learning_rate": 3e-06,
+      "loss": -0.0525,
+      "policy/loss": 1.3377684737747586e-07,
+      "step": 278
+    },
+    {
+      "epoch": 0.1488,
+      "grad_norm": 0.1881897896528244,
+      "learning_rate": 3e-06,
+      "loss": -0.0257,
+      "policy/loss": 1.2082745826091923e-07,
+      "step": 279
+    },
+    {
+      "epoch": 0.14933333333333335,
+      "grad_norm": 0.2116159200668335,
+      "learning_rate": 3e-06,
+      "loss": -0.0529,
+      "policy/loss": 1.334738151115289e-07,
+      "step": 280
+    },
+    {
+      "completion_length": 243.21875762939453,
+      "epoch": 0.14986666666666668,
+      "grad_norm": 0.17542949318885803,
+      "learning_rate": 3e-06,
+      "loss": -0.0625,
+      "policy/loss": 0.23601744323968887,
+      "reward": 1.1510416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.4375,
+      "step": 281
+    },
+    {
+      "epoch": 0.1504,
+      "grad_norm": 0.17792236804962158,
+      "learning_rate": 3e-06,
+      "loss": 0.1157,
+      "policy/loss": 0.8459010720252991,
+      "step": 282
+    },
+    {
+      "epoch": 0.15093333333333334,
+      "grad_norm": 0.1947941780090332,
+      "learning_rate": 3e-06,
+      "loss": -0.064,
+      "policy/loss": 0.23519348353147507,
+      "step": 283
+    },
+    {
+      "epoch": 0.15146666666666667,
+      "grad_norm": 0.16937273740768433,
+      "learning_rate": 3e-06,
+      "loss": 0.1152,
+      "policy/loss": 0.847868949174881,
+      "step": 284
+    },
+    {
+      "completion_length": 230.05208587646484,
+      "epoch": 0.152,
+      "grad_norm": 0.19694648683071136,
+      "learning_rate": 3e-06,
+      "loss": -0.0426,
+      "policy/loss": -0.16977207361173896,
+      "reward": 1.3125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333134651184,
+      "rewards/correctness_reward_func_math": 0.6041666716337204,
+      "step": 285
+    },
+    {
+      "epoch": 0.15253333333333333,
+      "grad_norm": 0.19341500103473663,
+      "learning_rate": 3e-06,
+      "loss": 0.0504,
+      "policy/loss": 0.24198517929160346,
+      "step": 286
+    },
+    {
+      "epoch": 0.15306666666666666,
+      "grad_norm": 0.20095853507518768,
+      "learning_rate": 3e-06,
+      "loss": -0.0433,
+      "policy/loss": -0.16863638153844818,
+      "step": 287
+    },
+    {
+      "epoch": 0.1536,
+      "grad_norm": 0.19524559378623962,
+      "learning_rate": 3e-06,
+      "loss": 0.0501,
+      "policy/loss": 0.24204269780870646,
+      "step": 288
+    },
+    {
+      "completion_length": 236.30209350585938,
+      "epoch": 0.15413333333333334,
+      "grad_norm": 0.1793384999036789,
+      "learning_rate": 3e-06,
+      "loss": -0.0062,
+      "policy/loss": -0.050802123099309426,
+      "reward": 1.4166666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.6666666865348816,
+      "step": 289
+    },
+    {
+      "epoch": 0.15466666666666667,
+      "grad_norm": 0.15829868614673615,
+      "learning_rate": 3e-06,
+      "loss": -0.0323,
+      "policy/loss": 0.5184337169803612,
+      "step": 290
+    },
+    {
+      "epoch": 0.1552,
+      "grad_norm": 0.18136386573314667,
+      "learning_rate": 3e-06,
+      "loss": -0.0067,
+      "policy/loss": -0.05044223028183126,
+      "step": 291
+    },
+    {
+      "epoch": 0.15573333333333333,
+      "grad_norm": 0.16140909492969513,
+      "learning_rate": 3e-06,
+      "loss": -0.0332,
+      "policy/loss": 0.5172658833048018,
+      "step": 292
+    },
+    {
+      "completion_length": 242.84375762939453,
+      "epoch": 0.15626666666666666,
+      "grad_norm": 0.217057004570961,
+      "learning_rate": 3e-06,
+      "loss": -0.0577,
+      "policy/loss": -0.26913838636740195,
+      "reward": 1.53125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.875,
+      "step": 293
+    },
+    {
+      "epoch": 0.1568,
+      "grad_norm": 0.17376776039600372,
+      "learning_rate": 3e-06,
+      "loss": -0.0015,
+      "policy/loss": 0.14191918877971688,
+      "step": 294
+    },
+    {
+      "epoch": 0.15733333333333333,
+      "grad_norm": 0.230887308716774,
+      "learning_rate": 3e-06,
+      "loss": -0.0577,
+      "policy/loss": -0.26548798934049245,
+      "step": 295
+    },
+    {
+      "epoch": 0.15786666666666666,
+      "grad_norm": 0.18917903304100037,
+      "learning_rate": 3e-06,
+      "loss": -0.0023,
+      "policy/loss": 0.13973698820755232,
+      "step": 296
+    },
+    {
+      "completion_length": 234.9375,
+      "epoch": 0.1584,
+      "grad_norm": 0.2030014991760254,
+      "learning_rate": 3e-06,
+      "loss": 0.0852,
+      "policy/loss": -0.024586310229267383,
+      "reward": 1.0520833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.3958333432674408,
+      "step": 297
+    },
+    {
+      "epoch": 0.15893333333333334,
+      "grad_norm": 0.19274838268756866,
+      "learning_rate": 3e-06,
+      "loss": -0.0039,
+      "policy/loss": -0.26369657049921,
+      "step": 298
+    },
+    {
+      "epoch": 0.15946666666666667,
+      "grad_norm": 0.2258785516023636,
+      "learning_rate": 3e-06,
+      "loss": 0.0843,
+      "policy/loss": -0.02440718710658807,
+      "step": 299
+    },
+    {
+      "epoch": 0.16,
+      "grad_norm": 0.26021164655685425,
+      "learning_rate": 3e-06,
+      "loss": -0.0043,
+      "policy/loss": -0.265295763137118,
+      "step": 300
+    },
+    {
+      "completion_length": 228.1354217529297,
+      "epoch": 0.16053333333333333,
+      "grad_norm": 0.16252881288528442,
+      "learning_rate": 3e-06,
+      "loss": -0.0906,
+      "policy/loss": 9.862675653948827e-08,
+      "reward": 1.3281250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.703125,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 301
+    },
+    {
+      "epoch": 0.16106666666666666,
+      "grad_norm": 0.16389980912208557,
+      "learning_rate": 3e-06,
+      "loss": 0.0083,
+      "policy/loss": 8.158900755006471e-08,
+      "step": 302
+    },
+    {
+      "epoch": 0.1616,
+      "grad_norm": 0.1610758751630783,
+      "learning_rate": 3e-06,
+      "loss": -0.0921,
+      "policy/loss": 9.878370477167664e-08,
+      "step": 303
+    },
+    {
+      "epoch": 0.16213333333333332,
+      "grad_norm": 0.17360134422779083,
+      "learning_rate": 3e-06,
+      "loss": 0.0081,
+      "policy/loss": 8.154112052238816e-08,
+      "step": 304
+    },
+    {
+      "completion_length": 235.9479217529297,
+      "epoch": 0.16266666666666665,
+      "grad_norm": 0.19056282937526703,
+      "learning_rate": 3e-06,
+      "loss": -0.0627,
+      "policy/loss": -0.09038293361663818,
+      "reward": 1.3177083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6927083134651184,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 305
+    },
+    {
+      "epoch": 0.1632,
+      "grad_norm": 0.18738578259944916,
+      "learning_rate": 3e-06,
+      "loss": -0.0495,
+      "policy/loss": -0.01942083239555359,
+      "step": 306
+    },
+    {
+      "epoch": 0.16373333333333334,
+      "grad_norm": 0.18269182741641998,
+      "learning_rate": 3e-06,
+      "loss": -0.0622,
+      "policy/loss": -0.0906238853931427,
+      "step": 307
+    },
+    {
+      "epoch": 0.16426666666666667,
+      "grad_norm": 0.19301725924015045,
+      "learning_rate": 3e-06,
+      "loss": -0.0496,
+      "policy/loss": -0.015904545783996582,
+      "step": 308
+    },
+    {
+      "completion_length": 230.56250762939453,
+      "epoch": 0.1648,
+      "grad_norm": 0.3638012409210205,
+      "learning_rate": 3e-06,
+      "loss": 0.1576,
+      "policy/loss": 0.3832428579053655,
+      "reward": 1.3125000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6458333134651184,
+      "rewards/correctness_reward_func_math": 0.6666666567325592,
+      "step": 309
+    },
+    {
+      "epoch": 0.16533333333333333,
+      "grad_norm": 0.2560274004936218,
+      "learning_rate": 3e-06,
+      "loss": 0.0764,
+      "policy/loss": -0.2121401718994136,
+      "step": 310
+    },
+    {
+      "epoch": 0.16586666666666666,
+      "grad_norm": 0.29017674922943115,
+      "learning_rate": 3e-06,
+      "loss": 0.157,
+      "policy/loss": 0.3837818328849636,
+      "step": 311
+    },
+    {
+      "epoch": 0.1664,
+      "grad_norm": 0.23071832954883575,
+      "learning_rate": 3e-06,
+      "loss": 0.0767,
+      "policy/loss": -0.2122447482189429,
+      "step": 312
+    },
+    {
+      "completion_length": 239.46875762939453,
+      "epoch": 0.16693333333333332,
+      "grad_norm": 0.23338288068771362,
+      "learning_rate": 3e-06,
+      "loss": -0.0353,
+      "policy/loss": 0.33510033694481933,
+      "reward": 1.484375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.7708333432674408,
+      "step": 313
+    },
+    {
+      "epoch": 0.16746666666666668,
+      "grad_norm": 0.21438926458358765,
+      "learning_rate": 3e-06,
+      "loss": -0.038,
+      "policy/loss": -0.17123721710945716,
+      "step": 314
+    },
+    {
+      "epoch": 0.168,
+      "grad_norm": 0.2394193410873413,
+      "learning_rate": 3e-06,
+      "loss": -0.0344,
+      "policy/loss": 0.337918265638109,
+      "step": 315
+    },
+    {
+      "epoch": 0.16853333333333334,
+      "grad_norm": 0.21120727062225342,
+      "learning_rate": 3e-06,
+      "loss": -0.0386,
+      "policy/loss": -0.17399594359910253,
+      "step": 316
+    },
+    {
+      "completion_length": 233.12500762939453,
+      "epoch": 0.16906666666666667,
+      "grad_norm": 0.21799977123737335,
+      "learning_rate": 3e-06,
+      "loss": -0.1622,
+      "policy/loss": 1.1384786091639398e-07,
+      "reward": 1.4531250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.7708333432674408,
+      "step": 317
+    },
+    {
+      "epoch": 0.1696,
+      "grad_norm": 0.1853318065404892,
+      "learning_rate": 3e-06,
+      "loss": 0.0583,
+      "policy/loss": 1.1506038077868652e-07,
+      "step": 318
+    },
+    {
+      "epoch": 0.17013333333333333,
+      "grad_norm": 0.3352256119251251,
+      "learning_rate": 3e-06,
+      "loss": -0.1628,
+      "policy/loss": 1.1391430732032859e-07,
+      "step": 319
+    },
+    {
+      "epoch": 0.17066666666666666,
+      "grad_norm": 0.1795516014099121,
+      "learning_rate": 3e-06,
+      "loss": 0.0577,
+      "policy/loss": 1.1503071206675486e-07,
+      "step": 320
+    },
+    {
+      "completion_length": 222.36458587646484,
+      "epoch": 0.1712,
+      "grad_norm": 0.2438187450170517,
+      "learning_rate": 3e-06,
+      "loss": 0.0936,
+      "policy/loss": 1.634569457564794e-07,
+      "reward": 1.2968750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.5833333283662796,
+      "step": 321
+    },
+    {
+      "epoch": 0.17173333333333332,
+      "grad_norm": 0.34096208214759827,
+      "learning_rate": 3e-06,
+      "loss": -0.0069,
+      "policy/loss": 2.0071702522272972e-07,
+      "step": 322
+    },
+    {
+      "epoch": 0.17226666666666668,
+      "grad_norm": 0.21931925415992737,
+      "learning_rate": 3e-06,
+      "loss": 0.0935,
+      "policy/loss": 1.6340089814548264e-07,
+      "step": 323
+    },
+    {
+      "epoch": 0.1728,
+      "grad_norm": 0.20934435725212097,
+      "learning_rate": 3e-06,
+      "loss": -0.008,
+      "policy/loss": 2.0084311103119035e-07,
+      "step": 324
+    },
+    {
+      "completion_length": 240.59375,
+      "epoch": 0.17333333333333334,
+      "grad_norm": 0.17476677894592285,
+      "learning_rate": 3e-06,
+      "loss": -0.0013,
+      "policy/loss": 1.4731805109136076e-07,
+      "reward": 1.234375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.5208333432674408,
+      "step": 325
+    },
+    {
+      "epoch": 0.17386666666666667,
+      "grad_norm": 0.1717752069234848,
+      "learning_rate": 3e-06,
+      "loss": -0.0073,
+      "policy/loss": 1.901416055716254e-07,
+      "step": 326
+    },
+    {
+      "epoch": 0.1744,
+      "grad_norm": 0.16229568421840668,
+      "learning_rate": 3e-06,
+      "loss": -0.0021,
+      "policy/loss": 1.4738501974420615e-07,
+      "step": 327
+    },
+    {
+      "epoch": 0.17493333333333333,
+      "grad_norm": 0.1914823353290558,
+      "learning_rate": 3e-06,
+      "loss": -0.0086,
+      "policy/loss": 1.9037534570998105e-07,
+      "step": 328
+    },
+    {
+      "completion_length": 224.37500762939453,
+      "epoch": 0.17546666666666666,
+      "grad_norm": 0.2121347039937973,
+      "learning_rate": 3e-06,
+      "loss": -0.2144,
+      "policy/loss": -0.7999960482120514,
+      "reward": 1.166666716337204,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.5416666716337204,
+      "step": 329
+    },
+    {
+      "epoch": 0.176,
+      "grad_norm": 0.24915044009685516,
+      "learning_rate": 3e-06,
+      "loss": 0.1358,
+      "policy/loss": 0.2794909179210663,
+      "step": 330
+    },
+    {
+      "epoch": 0.17653333333333332,
+      "grad_norm": 0.22942717373371124,
+      "learning_rate": 3e-06,
+      "loss": -0.2155,
+      "policy/loss": -0.8015126436948776,
+      "step": 331
+    },
+    {
+      "epoch": 0.17706666666666668,
+      "grad_norm": 0.2512389123439789,
+      "learning_rate": 3e-06,
+      "loss": 0.136,
+      "policy/loss": 0.2801401615142822,
+      "step": 332
+    },
+    {
+      "completion_length": 235.92708587646484,
+      "epoch": 0.1776,
+      "grad_norm": 0.23643840849399567,
+      "learning_rate": 3e-06,
+      "loss": 0.0885,
+      "policy/loss": -0.12049439122210792,
+      "reward": 1.359375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 333
+    },
+    {
+      "epoch": 0.17813333333333334,
+      "grad_norm": 0.20423255860805511,
+      "learning_rate": 3e-06,
+      "loss": 0.2084,
+      "policy/loss": 0.11749207413505225,
+      "step": 334
+    },
+    {
+      "epoch": 0.17866666666666667,
+      "grad_norm": 0.2348330169916153,
+      "learning_rate": 3e-06,
+      "loss": 0.0883,
+      "policy/loss": -0.12300269621023574,
+      "step": 335
+    },
+    {
+      "epoch": 0.1792,
+      "grad_norm": 0.22016537189483643,
+      "learning_rate": 3e-06,
+      "loss": 0.2079,
+      "policy/loss": 0.11810970493400497,
+      "step": 336
+    },
+    {
+      "completion_length": 244.62500762939453,
+      "epoch": 0.17973333333333333,
+      "grad_norm": 0.2049584835767746,
+      "learning_rate": 3e-06,
+      "loss": -0.015,
+      "policy/loss": 6.489533710407613e-08,
+      "reward": 1.1250000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.4375000149011612,
+      "step": 337
+    },
+    {
+      "epoch": 0.18026666666666666,
+      "grad_norm": 0.1346663236618042,
+      "learning_rate": 3e-06,
+      "loss": -0.041,
+      "policy/loss": 6.478083847127891e-08,
+      "step": 338
+    },
+    {
+      "epoch": 0.1808,
+      "grad_norm": 0.18277087807655334,
+      "learning_rate": 3e-06,
+      "loss": -0.0148,
+      "policy/loss": 6.489469939197079e-08,
+      "step": 339
+    },
+    {
+      "epoch": 0.18133333333333335,
+      "grad_norm": 0.13341772556304932,
+      "learning_rate": 3e-06,
+      "loss": -0.0414,
+      "policy/loss": 6.486085624146654e-08,
+      "step": 340
+    },
+    {
+      "completion_length": 225.58333587646484,
+      "epoch": 0.18186666666666668,
+      "grad_norm": 0.22087866067886353,
+      "learning_rate": 3e-06,
+      "loss": 0.0153,
+      "policy/loss": 1.0230001556976731e-07,
+      "reward": 1.0937500298023224,
+      "rewards/boxed_and_answer_tags_format_reward": 0.59375,
+      "rewards/correctness_reward_func_math": 0.4999999850988388,
+      "step": 341
+    },
+    {
+      "epoch": 0.1824,
+      "grad_norm": 0.2614704668521881,
+      "learning_rate": 3e-06,
+      "loss": 0.0384,
+      "policy/loss": 9.911980747290272e-08,
+      "step": 342
+    },
+    {
+      "epoch": 0.18293333333333334,
+      "grad_norm": 0.21058595180511475,
+      "learning_rate": 3e-06,
+      "loss": 0.0145,
+      "policy/loss": 1.0220088597634458e-07,
+      "step": 343
+    },
+    {
+      "epoch": 0.18346666666666667,
+      "grad_norm": 0.2594062089920044,
+      "learning_rate": 3e-06,
+      "loss": 0.0373,
+      "policy/loss": 9.908657361279438e-08,
+      "step": 344
+    },
+    {
+      "completion_length": 240.0729217529297,
+      "epoch": 0.184,
+      "grad_norm": 0.1981377899646759,
+      "learning_rate": 3e-06,
+      "loss": -0.015,
+      "policy/loss": 1.2303640506416968e-07,
+      "reward": 1.2760416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.5625,
+      "step": 345
+    },
+    {
+      "epoch": 0.18453333333333333,
+      "grad_norm": 0.29948103427886963,
+      "learning_rate": 3e-06,
+      "loss": -0.028,
+      "policy/loss": 1.0078089118792377e-07,
+      "step": 346
+    },
+    {
+      "epoch": 0.18506666666666666,
+      "grad_norm": 0.19000372290611267,
+      "learning_rate": 3e-06,
+      "loss": -0.0155,
+      "policy/loss": 1.2300465535020066e-07,
+      "step": 347
+    },
+    {
+      "epoch": 0.1856,
+      "grad_norm": 0.2217448204755783,
+      "learning_rate": 3e-06,
+      "loss": -0.0291,
+      "policy/loss": 1.0074190637254787e-07,
+      "step": 348
+    },
+    {
+      "completion_length": 243.81250762939453,
+      "epoch": 0.18613333333333335,
+      "grad_norm": 0.38686323165893555,
+      "learning_rate": 3e-06,
+      "loss": -0.0939,
+      "policy/loss": -0.10636615939438343,
+      "reward": 1.2812500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.6250000149011612,
+      "step": 349
+    },
+    {
+      "epoch": 0.18666666666666668,
+      "grad_norm": 0.16630643606185913,
+      "learning_rate": 3e-06,
+      "loss": 0.1561,
+      "policy/loss": 0.01497996598482132,
+      "step": 350
+    },
+    {
+      "epoch": 0.1872,
+      "grad_norm": 0.20746314525604248,
+      "learning_rate": 3e-06,
+      "loss": -0.0943,
+      "policy/loss": -0.10720517113804817,
+      "step": 351
+    },
+    {
+      "epoch": 0.18773333333333334,
+      "grad_norm": 0.17379166185855865,
+      "learning_rate": 3e-06,
+      "loss": 0.1544,
+      "policy/loss": 0.008214622735977173,
+      "step": 352
+    },
+    {
+      "completion_length": 238.81250762939453,
+      "epoch": 0.18826666666666667,
+      "grad_norm": 0.23901595175266266,
+      "learning_rate": 3e-06,
+      "loss": 0.1003,
+      "policy/loss": 0.25580898922772377,
+      "reward": 1.2447916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.5000000149011612,
+      "step": 353
+    },
+    {
+      "epoch": 0.1888,
+      "grad_norm": 0.23660589754581451,
+      "learning_rate": 3e-06,
+      "loss": 0.1485,
+      "policy/loss": 0.5506345453392534,
+      "step": 354
+    },
+    {
+      "epoch": 0.18933333333333333,
+      "grad_norm": 0.25951698422431946,
+      "learning_rate": 3e-06,
+      "loss": 0.1002,
+      "policy/loss": 0.2529614368562676,
+      "step": 355
+    },
+    {
+      "epoch": 0.18986666666666666,
+      "grad_norm": 0.24150840938091278,
+      "learning_rate": 3e-06,
+      "loss": 0.1467,
+      "policy/loss": 0.5471353355659581,
+      "step": 356
+    },
+    {
+      "completion_length": 228.6979217529297,
+      "epoch": 0.1904,
+      "grad_norm": 0.24792945384979248,
+      "learning_rate": 3e-06,
+      "loss": 0.2011,
+      "policy/loss": 0.8702186644077301,
+      "reward": 1.229166716337204,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.6041666716337204,
+      "step": 357
+    },
+    {
+      "epoch": 0.19093333333333334,
+      "grad_norm": 0.21561826765537262,
+      "learning_rate": 3e-06,
+      "loss": 0.0245,
+      "policy/loss": -0.009782284498214722,
+      "step": 358
+    },
+    {
+      "epoch": 0.19146666666666667,
+      "grad_norm": 0.2629568874835968,
+      "learning_rate": 3e-06,
+      "loss": 0.2006,
+      "policy/loss": 0.8668336123228073,
+      "step": 359
+    },
+    {
+      "epoch": 0.192,
+      "grad_norm": 0.20394107699394226,
+      "learning_rate": 3e-06,
+      "loss": 0.0233,
+      "policy/loss": -0.006911665201187134,
+      "step": 360
+    },
+    {
+      "completion_length": 245.43750762939453,
+      "epoch": 0.19253333333333333,
+      "grad_norm": 0.1936134248971939,
+      "learning_rate": 3e-06,
+      "loss": -0.1514,
+      "policy/loss": 1.2281146055670433e-07,
+      "reward": 1.53125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6770833432674408,
+      "rewards/correctness_reward_func_math": 0.8541666567325592,
+      "step": 361
+    },
+    {
+      "epoch": 0.19306666666666666,
+      "grad_norm": 0.22812145948410034,
+      "learning_rate": 3e-06,
+      "loss": -0.0393,
+      "policy/loss": 1.3971355272701658e-07,
+      "step": 362
+    },
+    {
+      "epoch": 0.1936,
+      "grad_norm": 0.18327108025550842,
+      "learning_rate": 3e-06,
+      "loss": -0.1509,
+      "policy/loss": 1.2275335592448755e-07,
+      "step": 363
+    },
+    {
+      "epoch": 0.19413333333333332,
+      "grad_norm": 0.25876787304878235,
+      "learning_rate": 3e-06,
+      "loss": -0.0409,
+      "policy/loss": 1.3968885070880788e-07,
+      "step": 364
+    },
+    {
+      "completion_length": 245.73958587646484,
+      "epoch": 0.19466666666666665,
+      "grad_norm": 0.27671656012535095,
+      "learning_rate": 3e-06,
+      "loss": -0.0644,
+      "policy/loss": 1.0789469939709306e-07,
+      "reward": 1.6093750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.8958333134651184,
+      "step": 365
+    },
+    {
+      "epoch": 0.1952,
+      "grad_norm": 0.22265639901161194,
+      "learning_rate": 3e-06,
+      "loss": 0.0978,
+      "policy/loss": 1.0346443879427625e-07,
+      "step": 366
+    },
+    {
+      "epoch": 0.19573333333333334,
+      "grad_norm": 0.2968622148036957,
+      "learning_rate": 3e-06,
+      "loss": -0.0651,
+      "policy/loss": 1.0794944138581286e-07,
+      "step": 367
+    },
+    {
+      "epoch": 0.19626666666666667,
+      "grad_norm": 0.21419622004032135,
+      "learning_rate": 3e-06,
+      "loss": 0.0964,
+      "policy/loss": 1.0346468393152009e-07,
+      "step": 368
+    },
+    {
+      "completion_length": 237.15625,
+      "epoch": 0.1968,
+      "grad_norm": 0.19327084720134735,
+      "learning_rate": 3e-06,
+      "loss": 0.0146,
+      "policy/loss": -0.008604005505560508,
+      "reward": 1.4375000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7291666865348816,
+      "rewards/correctness_reward_func_math": 0.7083333432674408,
+      "step": 369
+    },
+    {
+      "epoch": 0.19733333333333333,
+      "grad_norm": 0.24212704598903656,
+      "learning_rate": 3e-06,
+      "loss": 0.0715,
+      "policy/loss": -0.0817370053424824,
+      "step": 370
+    },
+    {
+      "epoch": 0.19786666666666666,
+      "grad_norm": 0.2557942867279053,
+      "learning_rate": 3e-06,
+      "loss": 0.0144,
+      "policy/loss": -0.010046159255246323,
+      "step": 371
+    },
+    {
+      "epoch": 0.1984,
+      "grad_norm": 0.22302547097206116,
+      "learning_rate": 3e-06,
+      "loss": 0.0707,
+      "policy/loss": -0.08399093225434484,
+      "step": 372
+    },
+    {
+      "completion_length": 236.3125,
+      "epoch": 0.19893333333333332,
+      "grad_norm": 0.2276604026556015,
+      "learning_rate": 3e-06,
+      "loss": -0.0847,
+      "policy/loss": 1.720080788913947e-08,
+      "reward": 1.1354166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7395833432674408,
+      "rewards/correctness_reward_func_math": 0.3958333283662796,
+      "step": 373
+    },
+    {
+      "epoch": 0.19946666666666665,
+      "grad_norm": 0.22918841242790222,
+      "learning_rate": 3e-06,
+      "loss": -0.1141,
+      "policy/loss": 2.0052405957926567e-08,
+      "step": 374
+    },
+    {
+      "epoch": 0.2,
+      "grad_norm": 0.23410621285438538,
+      "learning_rate": 3e-06,
+      "loss": -0.0852,
+      "policy/loss": 1.7176599698132122e-08,
+      "step": 375
+    },
+    {
+      "epoch": 0.20053333333333334,
+      "grad_norm": 0.27192217111587524,
+      "learning_rate": 3e-06,
+      "loss": -0.1142,
+      "policy/loss": 2.0067746575591627e-08,
+      "step": 376
+    },
+    {
+      "completion_length": 243.03125,
+      "epoch": 0.20106666666666667,
+      "grad_norm": 0.22794723510742188,
+      "learning_rate": 3e-06,
+      "loss": -0.0684,
+      "policy/loss": -0.5014254655619865,
+      "reward": 1.4166666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 377
+    },
+    {
+      "epoch": 0.2016,
+      "grad_norm": 0.27352598309516907,
+      "learning_rate": 3e-06,
+      "loss": -0.0087,
+      "policy/loss": -0.21875375244324147,
+      "step": 378
+    },
+    {
+      "epoch": 0.20213333333333333,
+      "grad_norm": 0.21570146083831787,
+      "learning_rate": 3e-06,
+      "loss": -0.0692,
+      "policy/loss": -0.5041510069927888,
+      "step": 379
+    },
+    {
+      "epoch": 0.20266666666666666,
+      "grad_norm": 0.22530947625637054,
+      "learning_rate": 3e-06,
+      "loss": -0.0091,
+      "policy/loss": -0.21855231850053514,
+      "step": 380
+    },
+    {
+      "completion_length": 239.62500762939453,
+      "epoch": 0.2032,
+      "grad_norm": 0.18700046837329865,
+      "learning_rate": 3e-06,
+      "loss": 0.0048,
+      "policy/loss": 1.0334410127654792e-07,
+      "reward": 1.2760417461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 381
+    },
+    {
+      "epoch": 0.20373333333333332,
+      "grad_norm": 0.17587722837924957,
+      "learning_rate": 3e-06,
+      "loss": -0.0479,
+      "policy/loss": 1.0717805309923278e-07,
+      "step": 382
+    },
+    {
+      "epoch": 0.20426666666666668,
+      "grad_norm": 0.17681998014450073,
+      "learning_rate": 3e-06,
+      "loss": 0.0046,
+      "policy/loss": 1.0324566090957887e-07,
+      "step": 383
+    },
+    {
+      "epoch": 0.2048,
+      "grad_norm": 0.19897598028182983,
+      "learning_rate": 3e-06,
+      "loss": -0.0486,
+      "policy/loss": 1.0715518428128235e-07,
+      "step": 384
+    },
+    {
+      "completion_length": 238.98959350585938,
+      "epoch": 0.20533333333333334,
+      "grad_norm": 0.8404969573020935,
+      "learning_rate": 3e-06,
+      "loss": -0.0618,
+      "policy/loss": 0.6622691481414407,
+      "reward": 1.5520833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6979166567325592,
+      "rewards/correctness_reward_func_math": 0.854166641831398,
+      "step": 385
+    },
+    {
+      "epoch": 0.20586666666666667,
+      "grad_norm": 0.23672257363796234,
+      "learning_rate": 3e-06,
+      "loss": -0.0228,
+      "policy/loss": 0.2696230173888239,
+      "step": 386
+    },
+    {
+      "epoch": 0.2064,
+      "grad_norm": 0.2536720335483551,
+      "learning_rate": 3e-06,
+      "loss": -0.0617,
+      "policy/loss": 0.6652498571454544,
+      "step": 387
+    },
+    {
+      "epoch": 0.20693333333333333,
+      "grad_norm": 0.2585340738296509,
+      "learning_rate": 3e-06,
+      "loss": -0.0229,
+      "policy/loss": 0.26942790155602125,
+      "step": 388
+    },
+    {
+      "completion_length": 220.78125762939453,
+      "epoch": 0.20746666666666666,
+      "grad_norm": 0.22579233348369598,
+      "learning_rate": 3e-06,
+      "loss": 0.0521,
+      "policy/loss": 0.08963917569551683,
+      "reward": 1.4739583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.640625,
+      "rewards/correctness_reward_func_math": 0.8333333134651184,
+      "step": 389
+    },
+    {
+      "epoch": 0.208,
+      "grad_norm": 0.2476857453584671,
+      "learning_rate": 3e-06,
+      "loss": 0.0403,
+      "policy/loss": 0.13885617263028394,
+      "step": 390
+    },
+    {
+      "epoch": 0.20853333333333332,
+      "grad_norm": 0.23572464287281036,
+      "learning_rate": 3e-06,
+      "loss": 0.0516,
+      "policy/loss": 0.08960221337321883,
+      "step": 391
+    },
+    {
+      "epoch": 0.20906666666666668,
+      "grad_norm": 0.27898290753364563,
+      "learning_rate": 3e-06,
+      "loss": 0.039,
+      "policy/loss": 0.1389177888650721,
+      "step": 392
+    },
+    {
+      "completion_length": 228.68750762939453,
+      "epoch": 0.2096,
+      "grad_norm": 0.23240920901298523,
+      "learning_rate": 3e-06,
+      "loss": 0.0653,
+      "policy/loss": 0.21865689032689417,
+      "reward": 1.4479166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 393
+    },
+    {
+      "epoch": 0.21013333333333334,
+      "grad_norm": 0.23569993674755096,
+      "learning_rate": 3e-06,
+      "loss": -0.0717,
+      "policy/loss": -0.3273223429086016,
+      "step": 394
+    },
+    {
+      "epoch": 0.21066666666666667,
+      "grad_norm": 0.2203858345746994,
+      "learning_rate": 3e-06,
+      "loss": 0.0638,
+      "policy/loss": 0.21424866495722839,
+      "step": 395
+    },
+    {
+      "epoch": 0.2112,
+      "grad_norm": 0.22818459570407867,
+      "learning_rate": 3e-06,
+      "loss": -0.0732,
+      "policy/loss": -0.32748959354339036,
+      "step": 396
+    },
+    {
+      "completion_length": 232.8541717529297,
+      "epoch": 0.21173333333333333,
+      "grad_norm": 0.16229206323623657,
+      "learning_rate": 3e-06,
+      "loss": 0.1102,
+      "policy/loss": 1.0023077123832991e-07,
+      "reward": 1.4270833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.7083333507180214,
+      "step": 397
+    },
+    {
+      "epoch": 0.21226666666666666,
+      "grad_norm": 0.24424533545970917,
+      "learning_rate": 3e-06,
+      "loss": -0.0155,
+      "policy/loss": 9.604277551034102e-08,
+      "step": 398
+    },
+    {
+      "epoch": 0.2128,
+      "grad_norm": 0.2116089016199112,
+      "learning_rate": 3e-06,
+      "loss": 0.1102,
+      "policy/loss": 1.0036516684408525e-07,
+      "step": 399
+    },
+    {
+      "epoch": 0.21333333333333335,
+      "grad_norm": 0.32357609272003174,
+      "learning_rate": 3e-06,
+      "loss": -0.0164,
+      "policy/loss": 9.591898830763057e-08,
+      "step": 400
+    },
+    {
+      "completion_length": 246.17708587646484,
+      "epoch": 0.21386666666666668,
+      "grad_norm": 0.22774499654769897,
+      "learning_rate": 3e-06,
+      "loss": -0.0312,
+      "policy/loss": 9.207033713209967e-08,
+      "reward": 1.2864583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.5416666865348816,
+      "step": 401
+    },
+    {
+      "epoch": 0.2144,
+      "grad_norm": 0.32067304849624634,
+      "learning_rate": 3e-06,
+      "loss": 0.0038,
+      "policy/loss": 8.223313940902699e-08,
+      "step": 402
+    },
+    {
+      "epoch": 0.21493333333333334,
+      "grad_norm": 0.23817503452301025,
+      "learning_rate": 3e-06,
+      "loss": -0.0319,
+      "policy/loss": 9.221905727940793e-08,
+      "step": 403
+    },
+    {
+      "epoch": 0.21546666666666667,
+      "grad_norm": 0.25587210059165955,
+      "learning_rate": 3e-06,
+      "loss": 0.0031,
+      "policy/loss": 8.213262603362637e-08,
+      "step": 404
+    },
+    {
+      "completion_length": 239.6875,
+      "epoch": 0.216,
+      "grad_norm": 0.2181871235370636,
+      "learning_rate": 3e-06,
+      "loss": -0.0339,
+      "policy/loss": 1.2053131825950913e-07,
+      "reward": 1.3541667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333432674408,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 405
+    },
+    {
+      "epoch": 0.21653333333333333,
+      "grad_norm": 0.2303449809551239,
+      "learning_rate": 3e-06,
+      "loss": -0.0745,
+      "policy/loss": 1.0803019279137516e-07,
+      "step": 406
+    },
+    {
+      "epoch": 0.21706666666666666,
+      "grad_norm": 0.22701787948608398,
+      "learning_rate": 3e-06,
+      "loss": -0.0348,
+      "policy/loss": 1.2048502995298804e-07,
+      "step": 407
+    },
+    {
+      "epoch": 0.2176,
+      "grad_norm": 0.22309185564517975,
+      "learning_rate": 3e-06,
+      "loss": -0.0758,
+      "policy/loss": 1.0803353944766059e-07,
+      "step": 408
+    },
+    {
+      "completion_length": 239.27083587646484,
+      "epoch": 0.21813333333333335,
+      "grad_norm": 0.18860410153865814,
+      "learning_rate": 3e-06,
+      "loss": 0.0006,
+      "policy/loss": 0.3046880066394806,
+      "reward": 1.84375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 1.125,
+      "step": 409
+    },
+    {
+      "epoch": 0.21866666666666668,
+      "grad_norm": 0.18225856125354767,
+      "learning_rate": 3e-06,
+      "loss": -0.0611,
+      "policy/loss": -0.8999187648296356,
+      "step": 410
+    },
+    {
+      "epoch": 0.2192,
+      "grad_norm": 0.2401667833328247,
+      "learning_rate": 3e-06,
+      "loss": 0.0002,
+      "policy/loss": 0.30353909730911255,
+      "step": 411
+    },
+    {
+      "epoch": 0.21973333333333334,
+      "grad_norm": 0.18748770654201508,
+      "learning_rate": 3e-06,
+      "loss": -0.0616,
+      "policy/loss": -0.9060285985469818,
+      "step": 412
+    },
+    {
+      "completion_length": 242.42708587646484,
+      "epoch": 0.22026666666666667,
+      "grad_norm": 0.169342502951622,
+      "learning_rate": 3e-06,
+      "loss": 0.0388,
+      "policy/loss": -0.3633980820522744,
+      "reward": 1.2656250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.5208333283662796,
+      "step": 413
+    },
+    {
+      "epoch": 0.2208,
+      "grad_norm": 0.23208259046077728,
+      "learning_rate": 3e-06,
+      "loss": -0.1212,
+      "policy/loss": -0.0506151722028676,
+      "step": 414
+    },
+    {
+      "epoch": 0.22133333333333333,
+      "grad_norm": 0.17472803592681885,
+      "learning_rate": 3e-06,
+      "loss": 0.0384,
+      "policy/loss": -0.3640569220808523,
+      "step": 415
+    },
+    {
+      "epoch": 0.22186666666666666,
+      "grad_norm": 0.2288508415222168,
+      "learning_rate": 3e-06,
+      "loss": -0.1211,
+      "policy/loss": -0.053919530804563465,
+      "step": 416
+    },
+    {
+      "completion_length": 245.1041717529297,
+      "epoch": 0.2224,
+      "grad_norm": 0.438856303691864,
+      "learning_rate": 3e-06,
+      "loss": -0.0438,
+      "policy/loss": 0.08694370090961456,
+      "reward": 1.234375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.5208333283662796,
+      "step": 417
+    },
+    {
+      "epoch": 0.22293333333333334,
+      "grad_norm": 0.21024340391159058,
+      "learning_rate": 3e-06,
+      "loss": 0.0441,
+      "policy/loss": -0.016962748020887375,
+      "step": 418
+    },
+    {
+      "epoch": 0.22346666666666667,
+      "grad_norm": 0.23452293872833252,
+      "learning_rate": 3e-06,
+      "loss": -0.0433,
+      "policy/loss": 0.08957107365131378,
+      "step": 419
+    },
+    {
+      "epoch": 0.224,
+      "grad_norm": 0.2497847080230713,
+      "learning_rate": 3e-06,
+      "loss": 0.0421,
+      "policy/loss": -0.01927708461880684,
+      "step": 420
+    },
+    {
+      "completion_length": 240.27083587646484,
+      "epoch": 0.22453333333333333,
+      "grad_norm": 0.18296854197978973,
+      "learning_rate": 3e-06,
+      "loss": -0.0977,
+      "policy/loss": -0.03817936033010483,
+      "reward": 1.0833333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.4583333432674408,
+      "step": 421
+    },
+    {
+      "epoch": 0.22506666666666666,
+      "grad_norm": 1.7822562456130981,
+      "learning_rate": 3e-06,
+      "loss": -0.0227,
+      "policy/loss": 0.10537185147404671,
+      "step": 422
+    },
+    {
+      "epoch": 0.2256,
+      "grad_norm": 0.17055243253707886,
+      "learning_rate": 3e-06,
+      "loss": -0.0978,
+      "policy/loss": -0.038423068821430206,
+      "step": 423
+    },
+    {
+      "epoch": 0.22613333333333333,
+      "grad_norm": 0.3759995400905609,
+      "learning_rate": 3e-06,
+      "loss": -0.0236,
+      "policy/loss": 0.10181506723165512,
+      "step": 424
+    },
+    {
+      "completion_length": 238.73958587646484,
+      "epoch": 0.22666666666666666,
+      "grad_norm": 0.19102062284946442,
+      "learning_rate": 3e-06,
+      "loss": -0.004,
+      "policy/loss": -0.4065638627925239,
+      "reward": 1.2291667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.5416666865348816,
+      "step": 425
+    },
+    {
+      "epoch": 0.2272,
+      "grad_norm": 0.23571139574050903,
+      "learning_rate": 3e-06,
+      "loss": -0.093,
+      "policy/loss": -0.11426132132049105,
+      "step": 426
+    },
+    {
+      "epoch": 0.22773333333333334,
+      "grad_norm": 0.2058352828025818,
+      "learning_rate": 3e-06,
+      "loss": -0.0037,
+      "policy/loss": -0.4054148341951773,
+      "step": 427
+    },
+    {
+      "epoch": 0.22826666666666667,
+      "grad_norm": 0.1925303041934967,
+      "learning_rate": 3e-06,
+      "loss": -0.0945,
+      "policy/loss": -0.11783543916669004,
+      "step": 428
+    },
+    {
+      "completion_length": 243.5416717529297,
+      "epoch": 0.2288,
+      "grad_norm": 0.22822710871696472,
+      "learning_rate": 3e-06,
+      "loss": -0.077,
+      "policy/loss": 0.19510336965322495,
+      "reward": 1.2916666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333134651184,
+      "rewards/correctness_reward_func_math": 0.5833333283662796,
+      "step": 429
+    },
+    {
+      "epoch": 0.22933333333333333,
+      "grad_norm": 0.24407804012298584,
+      "learning_rate": 3e-06,
+      "loss": -0.2169,
+      "policy/loss": -0.21330923587083817,
+      "step": 430
+    },
+    {
+      "epoch": 0.22986666666666666,
+      "grad_norm": 0.24678534269332886,
+      "learning_rate": 3e-06,
+      "loss": -0.0785,
+      "policy/loss": 0.19484785199165344,
+      "step": 431
+    },
+    {
+      "epoch": 0.2304,
+      "grad_norm": 0.19122204184532166,
+      "learning_rate": 3e-06,
+      "loss": -0.2174,
+      "policy/loss": -0.2142089530825615,
+      "step": 432
+    },
+    {
+      "completion_length": 240.0416717529297,
+      "epoch": 0.23093333333333332,
+      "grad_norm": 0.1858026087284088,
+      "learning_rate": 3e-06,
+      "loss": 0.0873,
+      "policy/loss": 0.19032977521419525,
+      "reward": 1.6302083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.671875,
+      "rewards/correctness_reward_func_math": 0.9583333134651184,
+      "step": 433
+    },
+    {
+      "epoch": 0.23146666666666665,
+      "grad_norm": 0.16684448719024658,
+      "learning_rate": 3e-06,
+      "loss": 0.1225,
+      "policy/loss": 0.3405742943286896,
+      "step": 434
+    },
+    {
+      "epoch": 0.232,
+      "grad_norm": 0.17547768354415894,
+      "learning_rate": 3e-06,
+      "loss": 0.0871,
+      "policy/loss": 0.1897440105676651,
+      "step": 435
+    },
+    {
+      "epoch": 0.23253333333333334,
+      "grad_norm": 0.17311470210552216,
+      "learning_rate": 3e-06,
+      "loss": 0.1224,
+      "policy/loss": 0.34117770195007324,
+      "step": 436
+    },
+    {
+      "completion_length": 237.7604217529297,
+      "epoch": 0.23306666666666667,
+      "grad_norm": 0.22772133350372314,
+      "learning_rate": 3e-06,
+      "loss": 0.071,
+      "policy/loss": 0.18031060762530515,
+      "reward": 1.1666666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6458333134651184,
+      "rewards/correctness_reward_func_math": 0.5208333432674408,
+      "step": 437
+    },
+    {
+      "epoch": 0.2336,
+      "grad_norm": 0.2167007327079773,
+      "learning_rate": 3e-06,
+      "loss": 0.0896,
+      "policy/loss": 0.15299368264112445,
+      "step": 438
+    },
+    {
+      "epoch": 0.23413333333333333,
+      "grad_norm": 0.22845038771629333,
+      "learning_rate": 3e-06,
+      "loss": 0.0706,
+      "policy/loss": 0.18113467166752173,
+      "step": 439
+    },
+    {
+      "epoch": 0.23466666666666666,
+      "grad_norm": 0.21673594415187836,
+      "learning_rate": 3e-06,
+      "loss": 0.0884,
+      "policy/loss": 0.15219030145578927,
+      "step": 440
+    },
+    {
+      "completion_length": 231.5729217529297,
+      "epoch": 0.2352,
+      "grad_norm": 0.19741858541965485,
+      "learning_rate": 3e-06,
+      "loss": 0.1549,
+      "policy/loss": 0.5328412364770756,
+      "reward": 1.2760416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.5624999850988388,
+      "step": 441
+    },
+    {
+      "epoch": 0.23573333333333332,
+      "grad_norm": 0.18155930936336517,
+      "learning_rate": 3e-06,
+      "loss": -0.1199,
+      "policy/loss": -0.5168094302844253,
+      "step": 442
+    },
+    {
+      "epoch": 0.23626666666666668,
+      "grad_norm": 0.1962336301803589,
+      "learning_rate": 3e-06,
+      "loss": 0.1543,
+      "policy/loss": 0.5312373350768347,
+      "step": 443
+    },
+    {
+      "epoch": 0.2368,
+      "grad_norm": 0.1737622618675232,
+      "learning_rate": 3e-06,
+      "loss": -0.1202,
+      "policy/loss": -0.5184760357513909,
+      "step": 444
+    },
+    {
+      "completion_length": 232.80209350585938,
+      "epoch": 0.23733333333333334,
+      "grad_norm": 0.16694821417331696,
+      "learning_rate": 3e-06,
+      "loss": 0.0128,
+      "policy/loss": 1.321492888450848e-07,
+      "reward": 1.541666716337204,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.854166641831398,
+      "step": 445
+    },
+    {
+      "epoch": 0.23786666666666667,
+      "grad_norm": 0.22202499210834503,
+      "learning_rate": 3e-06,
+      "loss": -0.0357,
+      "policy/loss": 1.255617565476541e-07,
+      "step": 446
+    },
+    {
+      "epoch": 0.2384,
+      "grad_norm": 0.1640511304140091,
+      "learning_rate": 3e-06,
+      "loss": 0.0118,
+      "policy/loss": 1.3208338955905674e-07,
+      "step": 447
+    },
+    {
+      "epoch": 0.23893333333333333,
+      "grad_norm": 0.1881989985704422,
+      "learning_rate": 3e-06,
+      "loss": -0.0373,
+      "policy/loss": 1.2554241735074356e-07,
+      "step": 448
+    },
+    {
+      "completion_length": 246.86458587646484,
+      "epoch": 0.23946666666666666,
+      "grad_norm": 0.2247757762670517,
+      "learning_rate": 3e-06,
+      "loss": 0.1038,
+      "policy/loss": 1.1373726493957292e-07,
+      "reward": 1.1979166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7395833134651184,
+      "rewards/correctness_reward_func_math": 0.4583333283662796,
+      "step": 449
+    },
+    {
+      "epoch": 0.24,
+      "grad_norm": 0.2674126625061035,
+      "learning_rate": 3e-06,
+      "loss": -0.1895,
+      "policy/loss": 1.2026696438738327e-07,
+      "step": 450
+    },
+    {
+      "epoch": 0.24053333333333332,
+      "grad_norm": 0.2204744666814804,
+      "learning_rate": 3e-06,
+      "loss": 0.102,
+      "policy/loss": 1.1363284002641194e-07,
+      "step": 451
+    },
+    {
+      "epoch": 0.24106666666666668,
+      "grad_norm": 0.2537151873111725,
+      "learning_rate": 3e-06,
+      "loss": -0.1901,
+      "policy/loss": 1.2019013340136553e-07,
+      "step": 452
+    },
+    {
+      "completion_length": 242.6041717529297,
+      "epoch": 0.2416,
+      "grad_norm": 0.16942152380943298,
+      "learning_rate": 3e-06,
+      "loss": -0.0924,
+      "policy/loss": -0.14160317554060242,
+      "reward": 1.5208333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6458333134651184,
+      "rewards/correctness_reward_func_math": 0.875,
+      "step": 453
+    },
+    {
+      "epoch": 0.24213333333333334,
+      "grad_norm": 0.20109575986862183,
+      "learning_rate": 3e-06,
+      "loss": 0.0343,
+      "policy/loss": 0.06072133309098149,
+      "step": 454
+    },
+    {
+      "epoch": 0.24266666666666667,
+      "grad_norm": 0.16130667924880981,
+      "learning_rate": 3e-06,
+      "loss": -0.092,
+      "policy/loss": -0.14291207854662247,
+      "step": 455
+    },
+    {
+      "epoch": 0.2432,
+      "grad_norm": 0.19251598417758942,
+      "learning_rate": 3e-06,
+      "loss": 0.0338,
+      "policy/loss": 0.06104762388313034,
+      "step": 456
+    },
+    {
+      "completion_length": 230.92708587646484,
+      "epoch": 0.24373333333333333,
+      "grad_norm": 0.1828054040670395,
+      "learning_rate": 3e-06,
+      "loss": -0.056,
+      "policy/loss": 2.0789234156382008e-07,
+      "reward": 1.3958333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7083333432674408,
+      "step": 457
+    },
+    {
+      "epoch": 0.24426666666666666,
+      "grad_norm": 0.17124108970165253,
+      "learning_rate": 3e-06,
+      "loss": 0.027,
+      "policy/loss": 2.0271610878808133e-07,
+      "step": 458
+    },
+    {
+      "epoch": 0.2448,
+      "grad_norm": 0.1991703361272812,
+      "learning_rate": 3e-06,
+      "loss": -0.0563,
+      "policy/loss": 2.078589034226752e-07,
+      "step": 459
+    },
+    {
+      "epoch": 0.24533333333333332,
+      "grad_norm": 0.16810566186904907,
+      "learning_rate": 3e-06,
+      "loss": 0.0258,
+      "policy/loss": 2.0283113144614617e-07,
+      "step": 460
+    },
+    {
+      "completion_length": 238.8541717529297,
+      "epoch": 0.24586666666666668,
+      "grad_norm": 0.17978668212890625,
+      "learning_rate": 3e-06,
+      "loss": -0.0574,
+      "policy/loss": 1.323191369806409e-07,
+      "reward": 1.5989583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.640625,
+      "rewards/correctness_reward_func_math": 0.9583333432674408,
+      "step": 461
+    },
+    {
+      "epoch": 0.2464,
+      "grad_norm": 0.19432879984378815,
+      "learning_rate": 3e-06,
+      "loss": 0.172,
+      "policy/loss": 1.2790192727152316e-07,
+      "step": 462
+    },
+    {
+      "epoch": 0.24693333333333334,
+      "grad_norm": 0.1721692979335785,
+      "learning_rate": 3e-06,
+      "loss": -0.0567,
+      "policy/loss": 1.322546836490801e-07,
+      "step": 463
+    },
+    {
+      "epoch": 0.24746666666666667,
+      "grad_norm": 0.19701918959617615,
+      "learning_rate": 3e-06,
+      "loss": 0.1718,
+      "policy/loss": 1.277725942827601e-07,
+      "step": 464
+    },
+    {
+      "completion_length": 249.17708587646484,
+      "epoch": 0.248,
+      "grad_norm": 0.16581466794013977,
+      "learning_rate": 3e-06,
+      "loss": -0.0508,
+      "policy/loss": -0.2753586654161424,
+      "reward": 1.5625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 465
+    },
+    {
+      "epoch": 0.24853333333333333,
+      "grad_norm": 0.2427285760641098,
+      "learning_rate": 3e-06,
+      "loss": 0.0592,
+      "policy/loss": 0.34941401638889147,
+      "step": 466
+    },
+    {
+      "epoch": 0.24906666666666666,
+      "grad_norm": 0.16409088671207428,
+      "learning_rate": 3e-06,
+      "loss": -0.0519,
+      "policy/loss": -0.28183125298380673,
+      "step": 467
+    },
+    {
+      "epoch": 0.2496,
+      "grad_norm": 0.2015465497970581,
+      "learning_rate": 3e-06,
+      "loss": 0.0592,
+      "policy/loss": 0.3493703857929553,
+      "step": 468
+    },
+    {
+      "completion_length": 240.27084350585938,
+      "epoch": 0.2501333333333333,
+      "grad_norm": 0.23742476105690002,
+      "learning_rate": 3e-06,
+      "loss": 0.0178,
+      "policy/loss": 9.04028460979589e-08,
+      "reward": 1.2864583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.6041666567325592,
+      "step": 469
+    },
+    {
+      "epoch": 0.25066666666666665,
+      "grad_norm": 0.15684302151203156,
+      "learning_rate": 3e-06,
+      "loss": -0.0945,
+      "policy/loss": 8.161903330972109e-08,
+      "step": 470
+    },
+    {
+      "epoch": 0.2512,
+      "grad_norm": 0.15071775019168854,
+      "learning_rate": 3e-06,
+      "loss": 0.0174,
+      "policy/loss": 9.028485159490174e-08,
+      "step": 471
+    },
+    {
+      "epoch": 0.2517333333333333,
+      "grad_norm": 0.16931508481502533,
+      "learning_rate": 3e-06,
+      "loss": -0.0948,
+      "policy/loss": 8.160085052111299e-08,
+      "step": 472
+    },
+    {
+      "completion_length": 241.75000762939453,
+      "epoch": 0.25226666666666664,
+      "grad_norm": 0.1807669997215271,
+      "learning_rate": 3e-06,
+      "loss": 0.0841,
+      "policy/loss": -0.04880030496499188,
+      "reward": 1.2135416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.5,
+      "step": 473
+    },
+    {
+      "epoch": 0.2528,
+      "grad_norm": 0.2843763828277588,
+      "learning_rate": 3e-06,
+      "loss": 0.0026,
+      "policy/loss": 0.14969329768524275,
+      "step": 474
+    },
+    {
+      "epoch": 0.25333333333333335,
+      "grad_norm": 0.24817919731140137,
+      "learning_rate": 3e-06,
+      "loss": 0.0836,
+      "policy/loss": -0.05110159192867769,
+      "step": 475
+    },
+    {
+      "epoch": 0.2538666666666667,
+      "grad_norm": 0.2903302311897278,
+      "learning_rate": 3e-06,
+      "loss": 0.002,
+      "policy/loss": 0.15237067874872068,
+      "step": 476
+    },
+    {
+      "completion_length": 242.83333587646484,
+      "epoch": 0.2544,
+      "grad_norm": 0.26471447944641113,
+      "learning_rate": 3e-06,
+      "loss": 0.0226,
+      "policy/loss": -0.20570210861799154,
+      "reward": 1.2864583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.6041666567325592,
+      "step": 477
+    },
+    {
+      "epoch": 0.25493333333333335,
+      "grad_norm": 0.4160853326320648,
+      "learning_rate": 3e-06,
+      "loss": -0.1319,
+      "policy/loss": -0.593661946603028,
+      "step": 478
+    },
+    {
+      "epoch": 0.2554666666666667,
+      "grad_norm": 0.24955685436725616,
+      "learning_rate": 3e-06,
+      "loss": 0.0217,
+      "policy/loss": -0.2106692914770818,
+      "step": 479
+    },
+    {
+      "epoch": 0.256,
+      "grad_norm": 0.29642051458358765,
+      "learning_rate": 3e-06,
+      "loss": -0.1343,
+      "policy/loss": -0.6044094390539172,
+      "step": 480
+    },
+    {
+      "completion_length": 235.46875,
+      "epoch": 0.25653333333333334,
+      "grad_norm": 0.20831671357154846,
+      "learning_rate": 3e-06,
+      "loss": 0.0488,
+      "policy/loss": -0.024837004441241817,
+      "reward": 1.5208333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6458333432674408,
+      "rewards/correctness_reward_func_math": 0.875,
+      "step": 481
+    },
+    {
+      "epoch": 0.25706666666666667,
+      "grad_norm": 0.3059718906879425,
+      "learning_rate": 3e-06,
+      "loss": 0.008,
+      "policy/loss": 0.017986979949711213,
+      "step": 482
+    },
+    {
+      "epoch": 0.2576,
+      "grad_norm": 0.2898612320423126,
+      "learning_rate": 3e-06,
+      "loss": 0.0498,
+      "policy/loss": -0.025026552744577657,
+      "step": 483
+    },
+    {
+      "epoch": 0.2581333333333333,
+      "grad_norm": 0.27914804220199585,
+      "learning_rate": 3e-06,
+      "loss": 0.0084,
+      "policy/loss": 0.018063156846473305,
+      "step": 484
+    },
+    {
+      "completion_length": 232.80209350585938,
+      "epoch": 0.25866666666666666,
+      "grad_norm": 0.2842847406864166,
+      "learning_rate": 3e-06,
+      "loss": 0.0381,
+      "policy/loss": 0.20880059152841568,
+      "reward": 1.4895833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.7708333432674408,
+      "step": 485
+    },
+    {
+      "epoch": 0.2592,
+      "grad_norm": 0.24378131330013275,
+      "learning_rate": 3e-06,
+      "loss": 0.1484,
+      "policy/loss": 0.9093791544437408,
+      "step": 486
+    },
+    {
+      "epoch": 0.2597333333333333,
+      "grad_norm": 0.2776099443435669,
+      "learning_rate": 3e-06,
+      "loss": 0.0369,
+      "policy/loss": 0.2066025212407112,
+      "step": 487
+    },
+    {
+      "epoch": 0.26026666666666665,
+      "grad_norm": 0.21694928407669067,
+      "learning_rate": 3e-06,
+      "loss": 0.1482,
+      "policy/loss": 0.9038669168949127,
+      "step": 488
+    },
+    {
+      "completion_length": 237.86458587646484,
+      "epoch": 0.2608,
+      "grad_norm": 0.22000554203987122,
+      "learning_rate": 3e-06,
+      "loss": 0.0265,
+      "policy/loss": 1.3834745260510317e-07,
+      "reward": 1.2083333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.4583333283662796,
+      "step": 489
+    },
+    {
+      "epoch": 0.2613333333333333,
+      "grad_norm": 0.18821260333061218,
+      "learning_rate": 3e-06,
+      "loss": -0.0758,
+      "policy/loss": 1.1957874335166707e-07,
+      "step": 490
+    },
+    {
+      "epoch": 0.2618666666666667,
+      "grad_norm": 0.21370434761047363,
+      "learning_rate": 3e-06,
+      "loss": 0.0238,
+      "policy/loss": 1.3811544974373646e-07,
+      "step": 491
+    },
+    {
+      "epoch": 0.2624,
+      "grad_norm": 0.18872219324111938,
+      "learning_rate": 3e-06,
+      "loss": -0.076,
+      "policy/loss": 1.196884724663505e-07,
+      "step": 492
+    },
+    {
+      "completion_length": 231.5416717529297,
+      "epoch": 0.26293333333333335,
+      "grad_norm": 0.19147855043411255,
+      "learning_rate": 3e-06,
+      "loss": -0.0437,
+      "policy/loss": -0.18969652637720813,
+      "reward": 1.5468750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.8333333432674408,
+      "step": 493
+    },
+    {
+      "epoch": 0.2634666666666667,
+      "grad_norm": 0.22441226243972778,
+      "learning_rate": 3e-06,
+      "loss": 0.0767,
+      "policy/loss": 0.1258547547619564,
+      "step": 494
+    },
+    {
+      "epoch": 0.264,
+      "grad_norm": 0.1942484825849533,
+      "learning_rate": 3e-06,
+      "loss": -0.0445,
+      "policy/loss": -0.18793111130011297,
+      "step": 495
+    },
+    {
+      "epoch": 0.26453333333333334,
+      "grad_norm": 0.22000983357429504,
+      "learning_rate": 3e-06,
+      "loss": 0.0752,
+      "policy/loss": 0.12096105695003878,
+      "step": 496
+    },
+    {
+      "completion_length": 244.5729217529297,
+      "epoch": 0.2650666666666667,
+      "grad_norm": 0.2864699065685272,
+      "learning_rate": 3e-06,
+      "loss": -0.1192,
+      "policy/loss": -0.5510639697313309,
+      "reward": 1.3958333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7083333134651184,
+      "step": 497
+    },
+    {
+      "epoch": 0.2656,
+      "grad_norm": 0.19913476705551147,
+      "learning_rate": 3e-06,
+      "loss": 0.0175,
+      "policy/loss": 0.006882995367050171,
+      "step": 498
+    },
+    {
+      "epoch": 0.26613333333333333,
+      "grad_norm": 0.2842441201210022,
+      "learning_rate": 3e-06,
+      "loss": -0.1198,
+      "policy/loss": -0.550665557384491,
+      "step": 499
+    },
+    {
+      "epoch": 0.26666666666666666,
+      "grad_norm": 0.1992996782064438,
+      "learning_rate": 3e-06,
+      "loss": 0.0156,
+      "policy/loss": 0.003917068243026733,
+      "step": 500
+    },
+    {
+      "completion_length": 241.45834350585938,
+      "epoch": 0.2672,
+      "grad_norm": 0.24095101654529572,
+      "learning_rate": 3e-06,
+      "loss": 0.0617,
+      "policy/loss": -0.03494505838506612,
+      "reward": 1.0520833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6770833432674408,
+      "rewards/correctness_reward_func_math": 0.375,
+      "step": 501
+    },
+    {
+      "epoch": 0.2677333333333333,
+      "grad_norm": 0.24200020730495453,
+      "learning_rate": 3e-06,
+      "loss": 0.1244,
+      "policy/loss": 0.034944935156055124,
+      "step": 502
+    },
+    {
+      "epoch": 0.26826666666666665,
+      "grad_norm": 0.23801825940608978,
+      "learning_rate": 3e-06,
+      "loss": 0.0611,
+      "policy/loss": -0.034321716668401336,
+      "step": 503
+    },
+    {
+      "epoch": 0.2688,
+      "grad_norm": 0.2573195695877075,
+      "learning_rate": 3e-06,
+      "loss": 0.1246,
+      "policy/loss": 0.03573268508642258,
+      "step": 504
+    },
+    {
+      "completion_length": 248.5625,
+      "epoch": 0.2693333333333333,
+      "grad_norm": 0.1538291722536087,
+      "learning_rate": 3e-06,
+      "loss": -0.0305,
+      "policy/loss": 7.13524777040675e-08,
+      "reward": 1.4010417461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.6874999850988388,
+      "step": 505
+    },
+    {
+      "epoch": 0.26986666666666664,
+      "grad_norm": 0.16491347551345825,
+      "learning_rate": 3e-06,
+      "loss": 0.067,
+      "policy/loss": 9.03844110666796e-08,
+      "step": 506
+    },
+    {
+      "epoch": 0.2704,
+      "grad_norm": 0.1439589411020279,
+      "learning_rate": 3e-06,
+      "loss": -0.0313,
+      "policy/loss": 7.142547886473949e-08,
+      "step": 507
+    },
+    {
+      "epoch": 0.27093333333333336,
+      "grad_norm": 0.17203491926193237,
+      "learning_rate": 3e-06,
+      "loss": 0.0652,
+      "policy/loss": 9.032816450371683e-08,
+      "step": 508
+    },
+    {
+      "completion_length": 230.02083587646484,
+      "epoch": 0.2714666666666667,
+      "grad_norm": 0.1953967809677124,
+      "learning_rate": 3e-06,
+      "loss": -0.0422,
+      "policy/loss": -0.16858381567234204,
+      "reward": 1.6145833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.9583333432674408,
+      "step": 509
+    },
+    {
+      "epoch": 0.272,
+      "grad_norm": 0.22444742918014526,
+      "learning_rate": 3e-06,
+      "loss": 0.0221,
+      "policy/loss": -0.051265810406221135,
+      "step": 510
+    },
+    {
+      "epoch": 0.27253333333333335,
+      "grad_norm": 0.3004179298877716,
+      "learning_rate": 3e-06,
+      "loss": -0.0425,
+      "policy/loss": -0.16988193039303212,
+      "step": 511
+    },
+    {
+      "epoch": 0.2730666666666667,
+      "grad_norm": 0.211281418800354,
+      "learning_rate": 3e-06,
+      "loss": 0.0206,
+      "policy/loss": -0.05282710553670711,
+      "step": 512
+    },
+    {
+      "completion_length": 235.71875,
+      "epoch": 0.2736,
+      "grad_norm": 0.22438064217567444,
+      "learning_rate": 3e-06,
+      "loss": -0.1077,
+      "policy/loss": -0.0919664506853124,
+      "reward": 1.6875000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333432674408,
+      "rewards/correctness_reward_func_math": 0.9791666567325592,
+      "step": 513
+    },
+    {
+      "epoch": 0.27413333333333334,
+      "grad_norm": 0.2189481258392334,
+      "learning_rate": 3e-06,
+      "loss": -0.1005,
+      "policy/loss": 0.053811804792673,
+      "step": 514
+    },
+    {
+      "epoch": 0.27466666666666667,
+      "grad_norm": 0.23293040692806244,
+      "learning_rate": 3e-06,
+      "loss": -0.1084,
+      "policy/loss": -0.09111569870481162,
+      "step": 515
+    },
+    {
+      "epoch": 0.2752,
+      "grad_norm": 0.23135444521903992,
+      "learning_rate": 3e-06,
+      "loss": -0.1015,
+      "policy/loss": 0.05482765414130242,
+      "step": 516
+    },
+    {
+      "completion_length": 244.58333587646484,
+      "epoch": 0.27573333333333333,
+      "grad_norm": 0.20323006808757782,
+      "learning_rate": 3e-06,
+      "loss": -0.0913,
+      "policy/loss": 0.07237387125480055,
+      "reward": 1.057291716337204,
+      "rewards/boxed_and_answer_tags_format_reward": 0.640625,
+      "rewards/correctness_reward_func_math": 0.4166666716337204,
+      "step": 517
+    },
+    {
+      "epoch": 0.27626666666666666,
+      "grad_norm": 0.23073336482048035,
+      "learning_rate": 3e-06,
+      "loss": 0.1687,
+      "policy/loss": 0.4265699766653661,
+      "step": 518
+    },
+    {
+      "epoch": 0.2768,
+      "grad_norm": 0.2233223170042038,
+      "learning_rate": 3e-06,
+      "loss": -0.0921,
+      "policy/loss": 0.06965768750550794,
+      "step": 519
+    },
+    {
+      "epoch": 0.2773333333333333,
+      "grad_norm": 0.23181813955307007,
+      "learning_rate": 3e-06,
+      "loss": 0.1669,
+      "policy/loss": 0.42428744658635864,
+      "step": 520
+    },
+    {
+      "completion_length": 233.46875,
+      "epoch": 0.27786666666666665,
+      "grad_norm": 0.19859309494495392,
+      "learning_rate": 3e-06,
+      "loss": -0.0845,
+      "policy/loss": 9.617231100378376e-08,
+      "reward": 1.3072917461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 521
+    },
+    {
+      "epoch": 0.2784,
+      "grad_norm": 0.16588932275772095,
+      "learning_rate": 3e-06,
+      "loss": -0.0806,
+      "policy/loss": 7.506767829568162e-08,
+      "step": 522
+    },
+    {
+      "epoch": 0.2789333333333333,
+      "grad_norm": 0.2613452076911926,
+      "learning_rate": 3e-06,
+      "loss": -0.0846,
+      "policy/loss": 9.63433244294265e-08,
+      "step": 523
+    },
+    {
+      "epoch": 0.27946666666666664,
+      "grad_norm": 0.1758737862110138,
+      "learning_rate": 3e-06,
+      "loss": -0.0824,
+      "policy/loss": 7.494377918249029e-08,
+      "step": 524
+    },
+    {
+      "completion_length": 241.46875762939453,
+      "epoch": 0.28,
+      "grad_norm": 0.26182135939598083,
+      "learning_rate": 3e-06,
+      "loss": 0.0432,
+      "policy/loss": 0.0031139403581619263,
+      "reward": 1.140625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.4583333432674408,
+      "step": 525
+    },
+    {
+      "epoch": 0.28053333333333336,
+      "grad_norm": 0.2553021013736725,
+      "learning_rate": 3e-06,
+      "loss": 0.1683,
+      "policy/loss": -0.008727588327019475,
+      "step": 526
+    },
+    {
+      "epoch": 0.2810666666666667,
+      "grad_norm": 0.26792725920677185,
+      "learning_rate": 3e-06,
+      "loss": 0.0429,
+      "policy/loss": 0.002528086304664612,
+      "step": 527
+    },
+    {
+      "epoch": 0.2816,
+      "grad_norm": 0.24978765845298767,
+      "learning_rate": 3e-06,
+      "loss": 0.1672,
+      "policy/loss": -0.011037004180252552,
+      "step": 528
+    },
+    {
+      "completion_length": 243.02084350585938,
+      "epoch": 0.28213333333333335,
+      "grad_norm": 0.15369071066379547,
+      "learning_rate": 3e-06,
+      "loss": -0.0421,
+      "policy/loss": 5.3416190581856426e-08,
+      "reward": 1.4479166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 529
+    },
+    {
+      "epoch": 0.2826666666666667,
+      "grad_norm": 0.16754868626594543,
+      "learning_rate": 3e-06,
+      "loss": 0.0609,
+      "policy/loss": 6.477198688514818e-08,
+      "step": 530
+    },
+    {
+      "epoch": 0.2832,
+      "grad_norm": 0.15279948711395264,
+      "learning_rate": 3e-06,
+      "loss": -0.0432,
+      "policy/loss": 5.344276843288753e-08,
+      "step": 531
+    },
+    {
+      "epoch": 0.28373333333333334,
+      "grad_norm": 0.17444705963134766,
+      "learning_rate": 3e-06,
+      "loss": 0.0595,
+      "policy/loss": 6.462999380119072e-08,
+      "step": 532
+    },
+    {
+      "completion_length": 235.96875762939453,
+      "epoch": 0.28426666666666667,
+      "grad_norm": 0.20894497632980347,
+      "learning_rate": 3e-06,
+      "loss": 0.0712,
+      "policy/loss": 1.0563718433331815e-07,
+      "reward": 1.3020833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.59375,
+      "rewards/correctness_reward_func_math": 0.7083333134651184,
+      "step": 533
+    },
+    {
+      "epoch": 0.2848,
+      "grad_norm": 0.19458335638046265,
+      "learning_rate": 3e-06,
+      "loss": 0.0519,
+      "policy/loss": 1.1758022466779039e-07,
+      "step": 534
+    },
+    {
+      "epoch": 0.2853333333333333,
+      "grad_norm": 0.2922618091106415,
+      "learning_rate": 3e-06,
+      "loss": 0.0704,
+      "policy/loss": 1.0554145646324287e-07,
+      "step": 535
+    },
+    {
+      "epoch": 0.28586666666666666,
+      "grad_norm": 0.22207331657409668,
+      "learning_rate": 3e-06,
+      "loss": 0.0517,
+      "policy/loss": 1.1751347983590676e-07,
+      "step": 536
+    },
+    {
+      "completion_length": 240.67709350585938,
+      "epoch": 0.2864,
+      "grad_norm": 0.19028104841709137,
+      "learning_rate": 3e-06,
+      "loss": -0.0677,
+      "policy/loss": 1.6177444805975938e-07,
+      "reward": 1.3125000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333134651184,
+      "rewards/correctness_reward_func_math": 0.6041666567325592,
+      "step": 537
+    },
+    {
+      "epoch": 0.2869333333333333,
+      "grad_norm": 0.1715867668390274,
+      "learning_rate": 3e-06,
+      "loss": 0.3021,
+      "policy/loss": 1.6804552416260776e-07,
+      "step": 538
+    },
+    {
+      "epoch": 0.28746666666666665,
+      "grad_norm": 0.19678060710430145,
+      "learning_rate": 3e-06,
+      "loss": -0.0685,
+      "policy/loss": 1.6169416028333217e-07,
+      "step": 539
+    },
+    {
+      "epoch": 0.288,
+      "grad_norm": 0.16509605944156647,
+      "learning_rate": 3e-06,
+      "loss": 0.3008,
+      "policy/loss": 1.6824333926024337e-07,
+      "step": 540
+    },
+    {
+      "completion_length": 242.8541717529297,
+      "epoch": 0.2885333333333333,
+      "grad_norm": 0.21473576128482819,
+      "learning_rate": 3e-06,
+      "loss": -0.0695,
+      "policy/loss": 9.730684169539927e-08,
+      "reward": 1.1666667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.4791666716337204,
+      "step": 541
+    },
+    {
+      "epoch": 0.2890666666666667,
+      "grad_norm": 0.1885623037815094,
+      "learning_rate": 3e-06,
+      "loss": -0.0096,
+      "policy/loss": 1.0523800142436812e-07,
+      "step": 542
+    },
+    {
+      "epoch": 0.2896,
+      "grad_norm": 0.20314373075962067,
+      "learning_rate": 3e-06,
+      "loss": -0.0709,
+      "policy/loss": 9.725095750923174e-08,
+      "step": 543
+    },
+    {
+      "epoch": 0.29013333333333335,
+      "grad_norm": 0.18794915080070496,
+      "learning_rate": 3e-06,
+      "loss": -0.0099,
+      "policy/loss": 1.0527963567596998e-07,
+      "step": 544
+    },
+    {
+      "completion_length": 246.58333587646484,
+      "epoch": 0.2906666666666667,
+      "grad_norm": 0.2013859897851944,
+      "learning_rate": 3e-06,
+      "loss": -0.1559,
+      "policy/loss": -0.6000539498514357,
+      "reward": 1.5625000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.8750000298023224,
+      "step": 545
+    },
+    {
+      "epoch": 0.2912,
+      "grad_norm": 0.19828177988529205,
+      "learning_rate": 3e-06,
+      "loss": 0.0292,
+      "policy/loss": 0.1629031735650157,
+      "step": 546
+    },
+    {
+      "epoch": 0.29173333333333334,
+      "grad_norm": 0.19953987002372742,
+      "learning_rate": 3e-06,
+      "loss": -0.1564,
+      "policy/loss": -0.6032809452602859,
+      "step": 547
+    },
+    {
+      "epoch": 0.2922666666666667,
+      "grad_norm": 0.19918911159038544,
+      "learning_rate": 3e-06,
+      "loss": 0.0281,
+      "policy/loss": 0.16286770888424495,
+      "step": 548
+    },
+    {
+      "completion_length": 244.9166717529297,
+      "epoch": 0.2928,
+      "grad_norm": 0.20692962408065796,
+      "learning_rate": 3e-06,
+      "loss": 0.0251,
+      "policy/loss": -0.20820922248542928,
+      "reward": 1.375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 549
+    },
+    {
+      "epoch": 0.29333333333333333,
+      "grad_norm": 0.20641833543777466,
+      "learning_rate": 3e-06,
+      "loss": 0.1371,
+      "policy/loss": 0.30451586809884645,
+      "step": 550
+    },
+    {
+      "epoch": 0.29386666666666666,
+      "grad_norm": 0.21400313079357147,
+      "learning_rate": 3e-06,
+      "loss": 0.0256,
+      "policy/loss": -0.20956656932295203,
+      "step": 551
+    },
+    {
+      "epoch": 0.2944,
+      "grad_norm": 0.2022261917591095,
+      "learning_rate": 3e-06,
+      "loss": 0.1366,
+      "policy/loss": 0.30177792864716935,
+      "step": 552
+    },
+    {
+      "completion_length": 231.05209350585938,
+      "epoch": 0.2949333333333333,
+      "grad_norm": 0.19336189329624176,
+      "learning_rate": 3e-06,
+      "loss": 0.0213,
+      "policy/loss": 0.10890545533265339,
+      "reward": 1.6354166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.9166666865348816,
+      "step": 553
+    },
+    {
+      "epoch": 0.29546666666666666,
+      "grad_norm": 0.2124696969985962,
+      "learning_rate": 3e-06,
+      "loss": 0.0368,
+      "policy/loss": 0.0552838354131282,
+      "step": 554
+    },
+    {
+      "epoch": 0.296,
+      "grad_norm": 0.19669975340366364,
+      "learning_rate": 3e-06,
+      "loss": 0.0201,
+      "policy/loss": 0.11037101430142471,
+      "step": 555
+    },
+    {
+      "epoch": 0.2965333333333333,
+      "grad_norm": 0.21314837038516998,
+      "learning_rate": 3e-06,
+      "loss": 0.0352,
+      "policy/loss": 0.05544924945594687,
+      "step": 556
+    },
+    {
+      "completion_length": 243.53125762939453,
+      "epoch": 0.29706666666666665,
+      "grad_norm": 0.20965850353240967,
+      "learning_rate": 3e-06,
+      "loss": 0.0247,
+      "policy/loss": 0.24432645738124847,
+      "reward": 1.4166666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 557
+    },
+    {
+      "epoch": 0.2976,
+      "grad_norm": 0.18962788581848145,
+      "learning_rate": 3e-06,
+      "loss": 0.1088,
+      "policy/loss": 0.17946629971265793,
+      "step": 558
+    },
+    {
+      "epoch": 0.2981333333333333,
+      "grad_norm": 0.2356860190629959,
+      "learning_rate": 3e-06,
+      "loss": 0.0238,
+      "policy/loss": 0.2461812049150467,
+      "step": 559
+    },
+    {
+      "epoch": 0.2986666666666667,
+      "grad_norm": 0.19349722564220428,
+      "learning_rate": 3e-06,
+      "loss": 0.1073,
+      "policy/loss": 0.1725727543234825,
+      "step": 560
+    },
+    {
+      "completion_length": 238.84375762939453,
+      "epoch": 0.2992,
+      "grad_norm": 0.3083341419696808,
+      "learning_rate": 3e-06,
+      "loss": -0.0292,
+      "policy/loss": 0.13254732596376329,
+      "reward": 1.1875000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.5000000149011612,
+      "step": 561
+    },
+    {
+      "epoch": 0.29973333333333335,
+      "grad_norm": 0.21715150773525238,
+      "learning_rate": 3e-06,
+      "loss": 0.0556,
+      "policy/loss": 0.339725020834738,
+      "step": 562
+    },
+    {
+      "epoch": 0.3002666666666667,
+      "grad_norm": 0.3062623143196106,
+      "learning_rate": 3e-06,
+      "loss": -0.0295,
+      "policy/loss": 0.1333444933457102,
+      "step": 563
+    },
+    {
+      "epoch": 0.3008,
+      "grad_norm": 0.21713200211524963,
+      "learning_rate": 3e-06,
+      "loss": 0.0553,
+      "policy/loss": 0.3377861115088514,
+      "step": 564
+    },
+    {
+      "completion_length": 244.96875,
+      "epoch": 0.30133333333333334,
+      "grad_norm": 0.25524815917015076,
+      "learning_rate": 3e-06,
+      "loss": -0.096,
+      "policy/loss": -0.3671819302342172,
+      "reward": 1.5572916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 565
+    },
+    {
+      "epoch": 0.30186666666666667,
+      "grad_norm": 0.2357453554868698,
+      "learning_rate": 3e-06,
+      "loss": -0.1014,
+      "policy/loss": -0.07160569121953841,
+      "step": 566
+    },
+    {
+      "epoch": 0.3024,
+      "grad_norm": 0.25050806999206543,
+      "learning_rate": 3e-06,
+      "loss": -0.0979,
+      "policy/loss": -0.3676633271396206,
+      "step": 567
+    },
+    {
+      "epoch": 0.30293333333333333,
+      "grad_norm": 0.22898314893245697,
+      "learning_rate": 3e-06,
+      "loss": -0.1033,
+      "policy/loss": -0.07265910650679785,
+      "step": 568
+    },
+    {
+      "completion_length": 240.6041717529297,
+      "epoch": 0.30346666666666666,
+      "grad_norm": 0.22358201444149017,
+      "learning_rate": 3e-06,
+      "loss": 0.0721,
+      "policy/loss": -0.20706608663889625,
+      "reward": 1.34375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 569
+    },
+    {
+      "epoch": 0.304,
+      "grad_norm": 0.21751290559768677,
+      "learning_rate": 3e-06,
+      "loss": 0.0848,
+      "policy/loss": -0.07522091302801215,
+      "step": 570
+    },
+    {
+      "epoch": 0.3045333333333333,
+      "grad_norm": 0.20164191722869873,
+      "learning_rate": 3e-06,
+      "loss": 0.071,
+      "policy/loss": -0.20801910522173017,
+      "step": 571
+    },
+    {
+      "epoch": 0.30506666666666665,
+      "grad_norm": 0.21564780175685883,
+      "learning_rate": 3e-06,
+      "loss": 0.0841,
+      "policy/loss": -0.07536664631818013,
+      "step": 572
+    },
+    {
+      "completion_length": 244.53125762939453,
+      "epoch": 0.3056,
+      "grad_norm": 0.2065366506576538,
+      "learning_rate": 3e-06,
+      "loss": 0.1275,
+      "policy/loss": 1.431055842715523e-07,
+      "reward": 1.390625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 573
+    },
+    {
+      "epoch": 0.3061333333333333,
+      "grad_norm": 0.2392921894788742,
+      "learning_rate": 3e-06,
+      "loss": 0.1315,
+      "policy/loss": 1.4050675645194133e-07,
+      "step": 574
+    },
+    {
+      "epoch": 0.30666666666666664,
+      "grad_norm": 0.2147788256406784,
+      "learning_rate": 3e-06,
+      "loss": 0.1271,
+      "policy/loss": 1.4302604256499762e-07,
+      "step": 575
+    },
+    {
+      "epoch": 0.3072,
+      "grad_norm": 0.23587584495544434,
+      "learning_rate": 3e-06,
+      "loss": 0.1315,
+      "policy/loss": 1.4057003028256077e-07,
+      "step": 576
+    },
+    {
+      "completion_length": 245.92708587646484,
+      "epoch": 0.30773333333333336,
+      "grad_norm": 0.28921079635620117,
+      "learning_rate": 3e-06,
+      "loss": -0.048,
+      "policy/loss": -0.14272324740886688,
+      "reward": 1.4739583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 577
+    },
+    {
+      "epoch": 0.3082666666666667,
+      "grad_norm": 0.2181992083787918,
+      "learning_rate": 3e-06,
+      "loss": 0.0422,
+      "policy/loss": 0.7535349279642105,
+      "step": 578
+    },
+    {
+      "epoch": 0.3088,
+      "grad_norm": 0.29032841324806213,
+      "learning_rate": 3e-06,
+      "loss": -0.0497,
+      "policy/loss": -0.15298626571893692,
+      "step": 579
+    },
+    {
+      "epoch": 0.30933333333333335,
+      "grad_norm": 0.2257576584815979,
+      "learning_rate": 3e-06,
+      "loss": 0.0408,
+      "policy/loss": 0.7442246899008751,
+      "step": 580
+    },
+    {
+      "completion_length": 233.7291717529297,
+      "epoch": 0.3098666666666667,
+      "grad_norm": 0.28347083926200867,
+      "learning_rate": 3e-06,
+      "loss": 0.1189,
+      "policy/loss": 0.09482800747947451,
+      "reward": 1.1250000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.5000000111758709,
+      "step": 581
+    },
+    {
+      "epoch": 0.3104,
+      "grad_norm": 0.15785036981105804,
+      "learning_rate": 3e-06,
+      "loss": 0.0509,
+      "policy/loss": 0.8142468932913189,
+      "step": 582
+    },
+    {
+      "epoch": 0.31093333333333334,
+      "grad_norm": 0.29626473784446716,
+      "learning_rate": 3e-06,
+      "loss": 0.1208,
+      "policy/loss": 0.09326921938901478,
+      "step": 583
+    },
+    {
+      "epoch": 0.31146666666666667,
+      "grad_norm": 0.15441589057445526,
+      "learning_rate": 3e-06,
+      "loss": 0.0504,
+      "policy/loss": 0.8116416338491206,
+      "step": 584
+    },
+    {
+      "completion_length": 239.6041717529297,
+      "epoch": 0.312,
+      "grad_norm": 0.16803021728992462,
+      "learning_rate": 3e-06,
+      "loss": -0.0576,
+      "policy/loss": 0.059079743660078066,
+      "reward": 1.5520833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.8958333134651184,
+      "step": 585
+    },
+    {
+      "epoch": 0.31253333333333333,
+      "grad_norm": 0.18079936504364014,
+      "learning_rate": 3e-06,
+      "loss": -0.0246,
+      "policy/loss": 0.0970631139437046,
+      "step": 586
+    },
+    {
+      "epoch": 0.31306666666666666,
+      "grad_norm": 0.1651124805212021,
+      "learning_rate": 3e-06,
+      "loss": -0.0578,
+      "policy/loss": 0.05841587461001296,
+      "step": 587
+    },
+    {
+      "epoch": 0.3136,
+      "grad_norm": 0.1789492964744568,
+      "learning_rate": 3e-06,
+      "loss": -0.0245,
+      "policy/loss": 0.09663966761306853,
+      "step": 588
+    },
+    {
+      "completion_length": 236.87500762939453,
+      "epoch": 0.3141333333333333,
+      "grad_norm": 0.24957488477230072,
+      "learning_rate": 3e-06,
+      "loss": 0.1296,
+      "policy/loss": -0.12270137696396688,
+      "reward": 1.2239583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6197916567325592,
+      "rewards/correctness_reward_func_math": 0.6041666865348816,
+      "step": 589
+    },
+    {
+      "epoch": 0.31466666666666665,
+      "grad_norm": 0.227695032954216,
+      "learning_rate": 3e-06,
+      "loss": -0.0546,
+      "policy/loss": -0.049846899642211895,
+      "step": 590
+    },
+    {
+      "epoch": 0.3152,
+      "grad_norm": 0.24034248292446136,
+      "learning_rate": 3e-06,
+      "loss": 0.1293,
+      "policy/loss": -0.12300947336672863,
+      "step": 591
+    },
+    {
+      "epoch": 0.3157333333333333,
+      "grad_norm": 0.21155031025409698,
+      "learning_rate": 3e-06,
+      "loss": -0.0544,
+      "policy/loss": -0.05011228943194368,
+      "step": 592
+    },
+    {
+      "completion_length": 244.06250762939453,
+      "epoch": 0.31626666666666664,
+      "grad_norm": 0.1803097128868103,
+      "learning_rate": 3e-06,
+      "loss": 0.1005,
+      "policy/loss": 1.1541287037175607e-07,
+      "reward": 1.4270833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.7708333432674408,
+      "step": 593
+    },
+    {
+      "epoch": 0.3168,
+      "grad_norm": 0.16887427866458893,
+      "learning_rate": 3e-06,
+      "loss": 0.0334,
+      "policy/loss": 1.1904116448135937e-07,
+      "step": 594
+    },
+    {
+      "epoch": 0.31733333333333336,
+      "grad_norm": 0.1847236156463623,
+      "learning_rate": 3e-06,
+      "loss": 0.1005,
+      "policy/loss": 1.1536621968843974e-07,
+      "step": 595
+    },
+    {
+      "epoch": 0.3178666666666667,
+      "grad_norm": 0.16868793964385986,
+      "learning_rate": 3e-06,
+      "loss": 0.0315,
+      "policy/loss": 1.1913300568267005e-07,
+      "step": 596
+    },
+    {
+      "completion_length": 238.2604217529297,
+      "epoch": 0.3184,
+      "grad_norm": 0.18538467586040497,
+      "learning_rate": 3e-06,
+      "loss": -0.0443,
+      "policy/loss": -0.22271379426967997,
+      "reward": 1.2447917461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.5625000149011612,
+      "step": 597
+    },
+    {
+      "epoch": 0.31893333333333335,
+      "grad_norm": 0.1698673814535141,
+      "learning_rate": 3e-06,
+      "loss": -0.0669,
+      "policy/loss": -0.4797472797998026,
+      "step": 598
+    },
+    {
+      "epoch": 0.3194666666666667,
+      "grad_norm": 0.19853878021240234,
+      "learning_rate": 3e-06,
+      "loss": -0.0445,
+      "policy/loss": -0.2234136570491927,
+      "step": 599
+    },
+    {
+      "epoch": 0.32,
+      "grad_norm": 0.1693759709596634,
+      "learning_rate": 3e-06,
+      "loss": -0.067,
+      "policy/loss": -0.48064903840383977,
+      "step": 600
+    },
+    {
+      "completion_length": 249.9166717529297,
+      "epoch": 0.32053333333333334,
+      "grad_norm": 0.24504069983959198,
+      "learning_rate": 3e-06,
+      "loss": 0.1668,
+      "policy/loss": 0.329084187746048,
+      "reward": 1.3333333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.5833333432674408,
+      "step": 601
+    },
+    {
+      "epoch": 0.32106666666666667,
+      "grad_norm": 0.17637869715690613,
+      "learning_rate": 3e-06,
+      "loss": -0.0914,
+      "policy/loss": -0.32428940385580063,
+      "step": 602
+    },
+    {
+      "epoch": 0.3216,
+      "grad_norm": 0.23216208815574646,
+      "learning_rate": 3e-06,
+      "loss": 0.1666,
+      "policy/loss": 0.3208381086587906,
+      "step": 603
+    },
+    {
+      "epoch": 0.3221333333333333,
+      "grad_norm": 0.1996132731437683,
+      "learning_rate": 3e-06,
+      "loss": -0.092,
+      "policy/loss": -0.3305753618478775,
+      "step": 604
+    },
+    {
+      "completion_length": 244.81250762939453,
+      "epoch": 0.32266666666666666,
+      "grad_norm": 0.199745774269104,
+      "learning_rate": 3e-06,
+      "loss": 0.0629,
+      "policy/loss": -0.007117834423496561,
+      "reward": 1.2968750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.5833333432674408,
+      "step": 605
+    },
+    {
+      "epoch": 0.3232,
+      "grad_norm": 0.1697131246328354,
+      "learning_rate": 3e-06,
+      "loss": -0.0279,
+      "policy/loss": 0.1074243944316855,
+      "step": 606
+    },
+    {
+      "epoch": 0.3237333333333333,
+      "grad_norm": 0.29605987668037415,
+      "learning_rate": 3e-06,
+      "loss": 0.0616,
+      "policy/loss": -0.007864481806475254,
+      "step": 607
+    },
+    {
+      "epoch": 0.32426666666666665,
+      "grad_norm": 0.18047109246253967,
+      "learning_rate": 3e-06,
+      "loss": -0.0282,
+      "policy/loss": 0.10894321762947712,
+      "step": 608
+    },
+    {
+      "completion_length": 237.4375,
+      "epoch": 0.3248,
+      "grad_norm": 0.17478826642036438,
+      "learning_rate": 3e-06,
+      "loss": -0.042,
+      "policy/loss": 1.476242701414776e-07,
+      "reward": 1.3385416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 609
+    },
+    {
+      "epoch": 0.3253333333333333,
+      "grad_norm": 0.2248871624469757,
+      "learning_rate": 3e-06,
+      "loss": -0.0068,
+      "policy/loss": 1.4220853827850988e-07,
+      "step": 610
+    },
+    {
+      "epoch": 0.3258666666666667,
+      "grad_norm": 0.17064714431762695,
+      "learning_rate": 3e-06,
+      "loss": -0.0423,
+      "policy/loss": 1.476768076713597e-07,
+      "step": 611
+    },
+    {
+      "epoch": 0.3264,
+      "grad_norm": 0.27181875705718994,
+      "learning_rate": 3e-06,
+      "loss": -0.0089,
+      "policy/loss": 1.4218650434827396e-07,
+      "step": 612
+    },
+    {
+      "completion_length": 249.18750762939453,
+      "epoch": 0.32693333333333335,
+      "grad_norm": 0.2307429313659668,
+      "learning_rate": 3e-06,
+      "loss": 0.1221,
+      "policy/loss": 0.05721418548303703,
+      "reward": 1.25,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.5625,
+      "step": 613
+    },
+    {
+      "epoch": 0.3274666666666667,
+      "grad_norm": 0.22401711344718933,
+      "learning_rate": 3e-06,
+      "loss": 0.0539,
+      "policy/loss": 0.19887694851033189,
+      "step": 614
+    },
+    {
+      "epoch": 0.328,
+      "grad_norm": 0.22318361699581146,
+      "learning_rate": 3e-06,
+      "loss": 0.1226,
+      "policy/loss": 0.05943179495891293,
+      "step": 615
+    },
+    {
+      "epoch": 0.32853333333333334,
+      "grad_norm": 0.2201337367296219,
+      "learning_rate": 3e-06,
+      "loss": 0.0532,
+      "policy/loss": 0.19921307400876032,
+      "step": 616
+    },
+    {
+      "completion_length": 247.45833587646484,
+      "epoch": 0.3290666666666667,
+      "grad_norm": 0.244893878698349,
+      "learning_rate": 3e-06,
+      "loss": 0.0449,
+      "policy/loss": 0.10334107664419889,
+      "reward": 1.4270833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7395833432674408,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 617
+    },
+    {
+      "epoch": 0.3296,
+      "grad_norm": 0.2506226599216461,
+      "learning_rate": 3e-06,
+      "loss": -0.0768,
+      "policy/loss": 0.08064769021873275,
+      "step": 618
+    },
+    {
+      "epoch": 0.33013333333333333,
+      "grad_norm": 0.25315216183662415,
+      "learning_rate": 3e-06,
+      "loss": 0.0445,
+      "policy/loss": 0.10276643575583222,
+      "step": 619
+    },
+    {
+      "epoch": 0.33066666666666666,
+      "grad_norm": 0.2550578713417053,
+      "learning_rate": 3e-06,
+      "loss": -0.0787,
+      "policy/loss": 0.07987629410140329,
+      "step": 620
+    },
+    {
+      "completion_length": 247.45834350585938,
+      "epoch": 0.3312,
+      "grad_norm": 0.2135765701532364,
+      "learning_rate": 3e-06,
+      "loss": -0.0107,
+      "policy/loss": 6.431333332557188e-08,
+      "reward": 1.3802083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 621
+    },
+    {
+      "epoch": 0.3317333333333333,
+      "grad_norm": 0.20224130153656006,
+      "learning_rate": 3e-06,
+      "loss": -0.0272,
+      "policy/loss": 7.163247062180744e-08,
+      "step": 622
+    },
+    {
+      "epoch": 0.33226666666666665,
+      "grad_norm": 0.20707233250141144,
+      "learning_rate": 3e-06,
+      "loss": -0.0111,
+      "policy/loss": 6.423586995651931e-08,
+      "step": 623
+    },
+    {
+      "epoch": 0.3328,
+      "grad_norm": 0.1732501983642578,
+      "learning_rate": 3e-06,
+      "loss": -0.0271,
+      "policy/loss": 7.168056015416369e-08,
+      "step": 624
+    },
+    {
+      "completion_length": 237.67708587646484,
+      "epoch": 0.3333333333333333,
+      "grad_norm": 0.2136276662349701,
+      "learning_rate": 3e-06,
+      "loss": 0.1712,
+      "policy/loss": -0.004264748241038063,
+      "reward": 1.3645833432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6145833134651184,
+      "rewards/correctness_reward_func_math": 0.7499999850988388,
+      "step": 625
+    },
+    {
+      "epoch": 0.33386666666666664,
+      "grad_norm": 0.1775953471660614,
+      "learning_rate": 3e-06,
+      "loss": -0.0621,
+      "policy/loss": -0.02116364061152254,
+      "step": 626
+    },
+    {
+      "epoch": 0.3344,
+      "grad_norm": 0.22402361035346985,
+      "learning_rate": 3e-06,
+      "loss": 0.1712,
+      "policy/loss": -0.004854573924362171,
+      "step": 627
+    },
+    {
+      "epoch": 0.33493333333333336,
+      "grad_norm": 0.17491792142391205,
+      "learning_rate": 3e-06,
+      "loss": -0.0634,
+      "policy/loss": -0.0218468067024844,
+      "step": 628
+    },
+    {
+      "completion_length": 244.1666717529297,
+      "epoch": 0.3354666666666667,
+      "grad_norm": 0.21437515318393707,
+      "learning_rate": 3e-06,
+      "loss": 0.1795,
+      "policy/loss": 1.6080770848247994e-07,
+      "reward": 1.4583333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7708333432674408,
+      "step": 629
+    },
+    {
+      "epoch": 0.336,
+      "grad_norm": 0.30284836888313293,
+      "learning_rate": 3e-06,
+      "loss": 0.1002,
+      "policy/loss": 1.4110529633626356e-07,
+      "step": 630
+    },
+    {
+      "epoch": 0.33653333333333335,
+      "grad_norm": 0.22314690053462982,
+      "learning_rate": 3e-06,
+      "loss": 0.1794,
+      "policy/loss": 1.607443493867322e-07,
+      "step": 631
+    },
+    {
+      "epoch": 0.3370666666666667,
+      "grad_norm": 0.29620814323425293,
+      "learning_rate": 3e-06,
+      "loss": 0.0999,
+      "policy/loss": 1.4105820866916474e-07,
+      "step": 632
+    },
+    {
+      "completion_length": 243.87500762939453,
+      "epoch": 0.3376,
+      "grad_norm": 0.19814875721931458,
+      "learning_rate": 3e-06,
+      "loss": -0.1154,
+      "policy/loss": -0.5281388163566589,
+      "reward": 1.6979166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.9791666567325592,
+      "step": 633
+    },
+    {
+      "epoch": 0.33813333333333334,
+      "grad_norm": 0.20505985617637634,
+      "learning_rate": 3e-06,
+      "loss": -0.1147,
+      "policy/loss": -0.08116238936781883,
+      "step": 634
+    },
+    {
+      "epoch": 0.33866666666666667,
+      "grad_norm": 0.1989133656024933,
+      "learning_rate": 3e-06,
+      "loss": -0.1146,
+      "policy/loss": -0.5262395143508911,
+      "step": 635
+    },
+    {
+      "epoch": 0.3392,
+      "grad_norm": 0.2205073982477188,
+      "learning_rate": 3e-06,
+      "loss": -0.1164,
+      "policy/loss": -0.08166669495403767,
+      "step": 636
+    },
+    {
+      "completion_length": 241.3541717529297,
+      "epoch": 0.33973333333333333,
+      "grad_norm": 0.18814793229103088,
+      "learning_rate": 3e-06,
+      "loss": -0.0039,
+      "policy/loss": 0.09649739414453506,
+      "reward": 1.3697917461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.6874999850988388,
+      "step": 637
+    },
+    {
+      "epoch": 0.34026666666666666,
+      "grad_norm": 0.2187618762254715,
+      "learning_rate": 3e-06,
+      "loss": -0.0239,
+      "policy/loss": 0.1784423291683197,
+      "step": 638
+    },
+    {
+      "epoch": 0.3408,
+      "grad_norm": 0.18895238637924194,
+      "learning_rate": 3e-06,
+      "loss": -0.005,
+      "policy/loss": 0.09430845826864243,
+      "step": 639
+    },
+    {
+      "epoch": 0.3413333333333333,
+      "grad_norm": 0.1991850733757019,
+      "learning_rate": 3e-06,
+      "loss": -0.0251,
+      "policy/loss": 0.17325830459594727,
+      "step": 640
+    },
+    {
+      "completion_length": 247.1041717529297,
+      "epoch": 0.34186666666666665,
+      "grad_norm": 0.20718678832054138,
+      "learning_rate": 3e-06,
+      "loss": 0.027,
+      "policy/loss": 6.21527078692452e-08,
+      "reward": 1.1979166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.4791666716337204,
+      "step": 641
+    },
+    {
+      "epoch": 0.3424,
+      "grad_norm": 0.19231341779232025,
+      "learning_rate": 3e-06,
+      "loss": -0.0757,
+      "policy/loss": 5.351958520805056e-08,
+      "step": 642
+    },
+    {
+      "epoch": 0.3429333333333333,
+      "grad_norm": 0.19835002720355988,
+      "learning_rate": 3e-06,
+      "loss": 0.0267,
+      "policy/loss": 6.216111891887977e-08,
+      "step": 643
+    },
+    {
+      "epoch": 0.34346666666666664,
+      "grad_norm": 0.19296002388000488,
+      "learning_rate": 3e-06,
+      "loss": -0.0774,
+      "policy/loss": 5.351267517994529e-08,
+      "step": 644
+    },
+    {
+      "completion_length": 245.1666717529297,
+      "epoch": 0.344,
+      "grad_norm": 0.18565228581428528,
+      "learning_rate": 3e-06,
+      "loss": 0.1792,
+      "policy/loss": 8.244640348209487e-08,
+      "reward": 1.5520833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8333333432674408,
+      "step": 645
+    },
+    {
+      "epoch": 0.34453333333333336,
+      "grad_norm": 0.18217000365257263,
+      "learning_rate": 3e-06,
+      "loss": -0.1423,
+      "policy/loss": 7.254468670225833e-08,
+      "step": 646
+    },
+    {
+      "epoch": 0.3450666666666667,
+      "grad_norm": 0.1872524470090866,
+      "learning_rate": 3e-06,
+      "loss": 0.1788,
+      "policy/loss": 8.239160287359937e-08,
+      "step": 647
+    },
+    {
+      "epoch": 0.3456,
+      "grad_norm": 0.19741125404834747,
+      "learning_rate": 3e-06,
+      "loss": -0.1432,
+      "policy/loss": 7.253462186440629e-08,
+      "step": 648
+    },
+    {
+      "completion_length": 224.71875762939453,
+      "epoch": 0.34613333333333335,
+      "grad_norm": 0.15290382504463196,
+      "learning_rate": 3e-06,
+      "loss": 0.0503,
+      "policy/loss": 0.010854335188895448,
+      "reward": 1.4739583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6197916567325592,
+      "rewards/correctness_reward_func_math": 0.8541666567325592,
+      "step": 649
+    },
+    {
+      "epoch": 0.3466666666666667,
+      "grad_norm": 0.1859896034002304,
+      "learning_rate": 3e-06,
+      "loss": -0.1092,
+      "policy/loss": -0.08057285803155878,
+      "step": 650
+    },
+    {
+      "epoch": 0.3472,
+      "grad_norm": 0.1624482423067093,
+      "learning_rate": 3e-06,
+      "loss": 0.0488,
+      "policy/loss": 0.005232166601672361,
+      "step": 651
+    },
+    {
+      "epoch": 0.34773333333333334,
+      "grad_norm": 0.18345311284065247,
+      "learning_rate": 3e-06,
+      "loss": -0.1095,
+      "policy/loss": -0.08235947741188454,
+      "step": 652
+    },
+    {
+      "completion_length": 234.6354217529297,
+      "epoch": 0.34826666666666667,
+      "grad_norm": 0.18288344144821167,
+      "learning_rate": 3e-06,
+      "loss": -0.1273,
+      "policy/loss": -0.1359227632851301,
+      "reward": 1.3072917461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.6249999850988388,
+      "step": 653
+    },
+    {
+      "epoch": 0.3488,
+      "grad_norm": 0.2629651427268982,
+      "learning_rate": 3e-06,
+      "loss": 0.0468,
+      "policy/loss": -0.1848104860528892,
+      "step": 654
+    },
+    {
+      "epoch": 0.34933333333333333,
+      "grad_norm": 0.20343472063541412,
+      "learning_rate": 3e-06,
+      "loss": -0.1281,
+      "policy/loss": -0.13793245299481072,
+      "step": 655
+    },
+    {
+      "epoch": 0.34986666666666666,
+      "grad_norm": 0.24522577226161957,
+      "learning_rate": 3e-06,
+      "loss": 0.0458,
+      "policy/loss": -0.18239138685473932,
+      "step": 656
+    },
+    {
+      "completion_length": 227.12500762939453,
+      "epoch": 0.3504,
+      "grad_norm": 0.20797525346279144,
+      "learning_rate": 3e-06,
+      "loss": 0.0305,
+      "policy/loss": 0.03603967032946542,
+      "reward": 1.2395833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6770833432674408,
+      "rewards/correctness_reward_func_math": 0.5625,
+      "step": 657
+    },
+    {
+      "epoch": 0.3509333333333333,
+      "grad_norm": 0.19227232038974762,
+      "learning_rate": 3e-06,
+      "loss": -0.092,
+      "policy/loss": 0.04128907413331717,
+      "step": 658
+    },
+    {
+      "epoch": 0.35146666666666665,
+      "grad_norm": 0.2379406839609146,
+      "learning_rate": 3e-06,
+      "loss": 0.0289,
+      "policy/loss": 0.03628325221833251,
+      "step": 659
+    },
+    {
+      "epoch": 0.352,
+      "grad_norm": 0.18634894490242004,
+      "learning_rate": 3e-06,
+      "loss": -0.0923,
+      "policy/loss": 0.041290027994620004,
+      "step": 660
+    },
+    {
+      "completion_length": 233.36458587646484,
+      "epoch": 0.3525333333333333,
+      "grad_norm": 0.21558566391468048,
+      "learning_rate": 3e-06,
+      "loss": -0.0544,
+      "policy/loss": -0.7696975469589233,
+      "reward": 1.6979166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.9791666567325592,
+      "step": 661
+    },
+    {
+      "epoch": 0.35306666666666664,
+      "grad_norm": 0.26031655073165894,
+      "learning_rate": 3e-06,
+      "loss": -0.1006,
+      "policy/loss": -0.058665379881858826,
+      "step": 662
+    },
+    {
+      "epoch": 0.3536,
+      "grad_norm": 0.22303546965122223,
+      "learning_rate": 3e-06,
+      "loss": -0.0546,
+      "policy/loss": -0.7698945999145508,
+      "step": 663
+    },
+    {
+      "epoch": 0.35413333333333336,
+      "grad_norm": 0.2694729268550873,
+      "learning_rate": 3e-06,
+      "loss": -0.1013,
+      "policy/loss": -0.06164480745792389,
+      "step": 664
+    },
+    {
+      "completion_length": 244.6354217529297,
+      "epoch": 0.3546666666666667,
+      "grad_norm": 0.18902617692947388,
+      "learning_rate": 3e-06,
+      "loss": 0.1019,
+      "policy/loss": 0.18437444644274237,
+      "reward": 1.1250000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.4999999850988388,
+      "step": 665
+    },
+    {
+      "epoch": 0.3552,
+      "grad_norm": 0.2124038189649582,
+      "learning_rate": 3e-06,
+      "loss": 0.1938,
+      "policy/loss": -0.16105667712983873,
+      "step": 666
+    },
+    {
+      "epoch": 0.35573333333333335,
+      "grad_norm": 0.18812984228134155,
+      "learning_rate": 3e-06,
+      "loss": 0.1019,
+      "policy/loss": 0.18322199058446031,
+      "step": 667
+    },
+    {
+      "epoch": 0.3562666666666667,
+      "grad_norm": 0.2215694785118103,
+      "learning_rate": 3e-06,
+      "loss": 0.1934,
+      "policy/loss": -0.16192021433396775,
+      "step": 668
+    },
+    {
+      "completion_length": 242.7604217529297,
+      "epoch": 0.3568,
+      "grad_norm": 0.19054600596427917,
+      "learning_rate": 3e-06,
+      "loss": -0.0015,
+      "policy/loss": 0.5612276191889478,
+      "reward": 1.0260416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.375,
+      "step": 669
+    },
+    {
+      "epoch": 0.35733333333333334,
+      "grad_norm": 0.27560585737228394,
+      "learning_rate": 3e-06,
+      "loss": 0.037,
+      "policy/loss": -0.4332189099323216,
+      "step": 670
+    },
+    {
+      "epoch": 0.35786666666666667,
+      "grad_norm": 0.1842101514339447,
+      "learning_rate": 3e-06,
+      "loss": -0.0021,
+      "policy/loss": 0.5611239070351246,
+      "step": 671
+    },
+    {
+      "epoch": 0.3584,
+      "grad_norm": 0.2113751918077469,
+      "learning_rate": 3e-06,
+      "loss": 0.0365,
+      "policy/loss": -0.43619523824231266,
+      "step": 672
+    },
+    {
+      "completion_length": 238.1354217529297,
+      "epoch": 0.3589333333333333,
+      "grad_norm": 0.24725672602653503,
+      "learning_rate": 3e-06,
+      "loss": -0.0023,
+      "policy/loss": 9.321555083374733e-08,
+      "reward": 1.3437500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 673
+    },
+    {
+      "epoch": 0.35946666666666666,
+      "grad_norm": 0.20894236862659454,
+      "learning_rate": 3e-06,
+      "loss": 0.1012,
+      "policy/loss": 1.0083192236720606e-07,
+      "step": 674
+    },
+    {
+      "epoch": 0.36,
+      "grad_norm": 0.2383037656545639,
+      "learning_rate": 3e-06,
+      "loss": -0.0024,
+      "policy/loss": 9.323236227487541e-08,
+      "step": 675
+    },
+    {
+      "epoch": 0.3605333333333333,
+      "grad_norm": 0.20646050572395325,
+      "learning_rate": 3e-06,
+      "loss": 0.1003,
+      "policy/loss": 1.0077053147483639e-07,
+      "step": 676
+    },
+    {
+      "completion_length": 230.59375762939453,
+      "epoch": 0.36106666666666665,
+      "grad_norm": 0.2313254028558731,
+      "learning_rate": 3e-06,
+      "loss": -0.0899,
+      "policy/loss": -0.01339622202912949,
+      "reward": 1.2552083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.6041666567325592,
+      "step": 677
+    },
+    {
+      "epoch": 0.3616,
+      "grad_norm": 0.2004411369562149,
+      "learning_rate": 3e-06,
+      "loss": 0.0106,
+      "policy/loss": -0.12382776187008915,
+      "step": 678
+    },
+    {
+      "epoch": 0.3621333333333333,
+      "grad_norm": 0.21355587244033813,
+      "learning_rate": 3e-06,
+      "loss": -0.0894,
+      "policy/loss": -0.011114417470423632,
+      "step": 679
+    },
+    {
+      "epoch": 0.3626666666666667,
+      "grad_norm": 0.19803400337696075,
+      "learning_rate": 3e-06,
+      "loss": 0.0101,
+      "policy/loss": -0.1230228756672247,
+      "step": 680
+    },
+    {
+      "completion_length": 244.8854217529297,
+      "epoch": 0.3632,
+      "grad_norm": 0.29710647463798523,
+      "learning_rate": 3e-06,
+      "loss": 0.1719,
+      "policy/loss": 0.00767888658121052,
+      "reward": 1.1354166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.4166666567325592,
+      "step": 681
+    },
+    {
+      "epoch": 0.36373333333333335,
+      "grad_norm": 0.24966536462306976,
+      "learning_rate": 3e-06,
+      "loss": 0.0671,
+      "policy/loss": -0.04807628439877476,
+      "step": 682
+    },
+    {
+      "epoch": 0.3642666666666667,
+      "grad_norm": 0.29531875252723694,
+      "learning_rate": 3e-06,
+      "loss": 0.1716,
+      "policy/loss": 0.00533518923320031,
+      "step": 683
+    },
+    {
+      "epoch": 0.3648,
+      "grad_norm": 0.24483022093772888,
+      "learning_rate": 3e-06,
+      "loss": 0.066,
+      "policy/loss": -0.048743078100230264,
+      "step": 684
+    },
+    {
+      "completion_length": 246.2604217529297,
+      "epoch": 0.36533333333333334,
+      "grad_norm": 0.22850942611694336,
+      "learning_rate": 3e-06,
+      "loss": 0.0886,
+      "policy/loss": 1.524917045969687e-07,
+      "reward": 1.4947917461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.7499999850988388,
+      "step": 685
+    },
+    {
+      "epoch": 0.3658666666666667,
+      "grad_norm": 0.22965452075004578,
+      "learning_rate": 3e-06,
+      "loss": 0.0027,
+      "policy/loss": 1.2643697644421081e-07,
+      "step": 686
+    },
+    {
+      "epoch": 0.3664,
+      "grad_norm": 0.24385984241962433,
+      "learning_rate": 3e-06,
+      "loss": 0.0894,
+      "policy/loss": 1.5258425634101513e-07,
+      "step": 687
+    },
+    {
+      "epoch": 0.36693333333333333,
+      "grad_norm": 0.25587567687034607,
+      "learning_rate": 3e-06,
+      "loss": 0.0018,
+      "policy/loss": 1.2646169267327423e-07,
+      "step": 688
+    },
+    {
+      "completion_length": 222.96875762939453,
+      "epoch": 0.36746666666666666,
+      "grad_norm": 0.21721050143241882,
+      "learning_rate": 3e-06,
+      "loss": -0.0128,
+      "policy/loss": 0.04586063373275451,
+      "reward": 1.2031250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6197916567325592,
+      "rewards/correctness_reward_func_math": 0.5833333134651184,
+      "step": 689
+    },
+    {
+      "epoch": 0.368,
+      "grad_norm": 0.17564712464809418,
+      "learning_rate": 3e-06,
+      "loss": -0.017,
+      "policy/loss": 0.046178413107480054,
+      "step": 690
+    },
+    {
+      "epoch": 0.3685333333333333,
+      "grad_norm": 0.22030027210712433,
+      "learning_rate": 3e-06,
+      "loss": -0.0124,
+      "policy/loss": 0.04571702377652542,
+      "step": 691
+    },
+    {
+      "epoch": 0.36906666666666665,
+      "grad_norm": 0.18415063619613647,
+      "learning_rate": 3e-06,
+      "loss": -0.0177,
+      "policy/loss": 0.04633554956507879,
+      "step": 692
+    },
+    {
+      "completion_length": 246.21875762939453,
+      "epoch": 0.3696,
+      "grad_norm": 0.15577328205108643,
+      "learning_rate": 3e-06,
+      "loss": 0.0276,
+      "policy/loss": 7.120588918496651e-08,
+      "reward": 1.0260416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.375,
+      "step": 693
+    },
+    {
+      "epoch": 0.3701333333333333,
+      "grad_norm": 0.14149676263332367,
+      "learning_rate": 3e-06,
+      "loss": -0.0586,
+      "policy/loss": 6.126570184505908e-08,
+      "step": 694
+    },
+    {
+      "epoch": 0.37066666666666664,
+      "grad_norm": 0.15049919486045837,
+      "learning_rate": 3e-06,
+      "loss": 0.0278,
+      "policy/loss": 7.12305094907606e-08,
+      "step": 695
+    },
+    {
+      "epoch": 0.3712,
+      "grad_norm": 0.13718090951442719,
+      "learning_rate": 3e-06,
+      "loss": -0.0586,
+      "policy/loss": 6.126876428425021e-08,
+      "step": 696
+    },
+    {
+      "completion_length": 247.17709350585938,
+      "epoch": 0.37173333333333336,
+      "grad_norm": 0.356935054063797,
+      "learning_rate": 3e-06,
+      "loss": -0.0002,
+      "policy/loss": 8.051970468159197e-08,
+      "reward": 1.2916666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.6041666641831398,
+      "step": 697
+    },
+    {
+      "epoch": 0.3722666666666667,
+      "grad_norm": 0.29510465264320374,
+      "learning_rate": 3e-06,
+      "loss": 0.1046,
+      "policy/loss": 8.383270788669961e-08,
+      "step": 698
+    },
+    {
+      "epoch": 0.3728,
+      "grad_norm": 0.2866463363170624,
+      "learning_rate": 3e-06,
+      "loss": -0.0021,
+      "policy/loss": 8.055803135675887e-08,
+      "step": 699
+    },
+    {
+      "epoch": 0.37333333333333335,
+      "grad_norm": 0.20645871758460999,
+      "learning_rate": 3e-06,
+      "loss": 0.1045,
+      "policy/loss": 8.379427107740867e-08,
+      "step": 700
+    },
+    {
+      "completion_length": 246.8541717529297,
+      "epoch": 0.3738666666666667,
+      "grad_norm": 0.3938835859298706,
+      "learning_rate": 3e-06,
+      "loss": -0.0003,
+      "policy/loss": -0.04606513378261212,
+      "reward": 1.4791666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 701
+    },
+    {
+      "epoch": 0.3744,
+      "grad_norm": 0.21100911498069763,
+      "learning_rate": 3e-06,
+      "loss": -0.0591,
+      "policy/loss": 0.15989056815905656,
+      "step": 702
+    },
+    {
+      "epoch": 0.37493333333333334,
+      "grad_norm": 0.2743366062641144,
+      "learning_rate": 3e-06,
+      "loss": -0.001,
+      "policy/loss": -0.05078388830504821,
+      "step": 703
+    },
+    {
+      "epoch": 0.37546666666666667,
+      "grad_norm": 0.2682023048400879,
+      "learning_rate": 3e-06,
+      "loss": -0.0605,
+      "policy/loss": 0.15512005068927692,
+      "step": 704
+    },
+    {
+      "completion_length": 239.7291717529297,
+      "epoch": 0.376,
+      "grad_norm": 0.20349057018756866,
+      "learning_rate": 3e-06,
+      "loss": -0.0274,
+      "policy/loss": 1.093363088955357e-07,
+      "reward": 1.328125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.5833333432674408,
+      "step": 705
+    },
+    {
+      "epoch": 0.37653333333333333,
+      "grad_norm": 0.25171977281570435,
+      "learning_rate": 3e-06,
+      "loss": -0.1212,
+      "policy/loss": 1.0154772311921079e-07,
+      "step": 706
+    },
+    {
+      "epoch": 0.37706666666666666,
+      "grad_norm": 0.1979297548532486,
+      "learning_rate": 3e-06,
+      "loss": -0.0283,
+      "policy/loss": 1.0916974702013249e-07,
+      "step": 707
+    },
+    {
+      "epoch": 0.3776,
+      "grad_norm": 0.26988741755485535,
+      "learning_rate": 3e-06,
+      "loss": -0.1216,
+      "policy/loss": 1.0152680118835633e-07,
+      "step": 708
+    },
+    {
+      "completion_length": 229.08333587646484,
+      "epoch": 0.3781333333333333,
+      "grad_norm": 0.21071836352348328,
+      "learning_rate": 3e-06,
+      "loss": -0.1178,
+      "policy/loss": -0.2186510433606692,
+      "reward": 1.859375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 1.1458333134651184,
+      "step": 709
+    },
+    {
+      "epoch": 0.37866666666666665,
+      "grad_norm": 0.21147918701171875,
+      "learning_rate": 3e-06,
+      "loss": -0.0471,
+      "policy/loss": 0.366060508198391,
+      "step": 710
+    },
+    {
+      "epoch": 0.3792,
+      "grad_norm": 0.20341691374778748,
+      "learning_rate": 3e-06,
+      "loss": -0.118,
+      "policy/loss": -0.21926264659950867,
+      "step": 711
+    },
+    {
+      "epoch": 0.3797333333333333,
+      "grad_norm": 0.16544830799102783,
+      "learning_rate": 3e-06,
+      "loss": -0.0476,
+      "policy/loss": 0.3661477991871358,
+      "step": 712
+    },
+    {
+      "completion_length": 235.65625762939453,
+      "epoch": 0.38026666666666664,
+      "grad_norm": 0.2719252407550812,
+      "learning_rate": 3e-06,
+      "loss": -0.1136,
+      "policy/loss": 2.0548662860164768e-07,
+      "reward": 1.3697916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.6874999850988388,
+      "step": 713
+    },
+    {
+      "epoch": 0.3808,
+      "grad_norm": 0.2634405493736267,
+      "learning_rate": 3e-06,
+      "loss": 0.2117,
+      "policy/loss": 2.0445702375582187e-07,
+      "step": 714
+    },
+    {
+      "epoch": 0.38133333333333336,
+      "grad_norm": 0.39286261796951294,
+      "learning_rate": 3e-06,
+      "loss": -0.1135,
+      "policy/loss": 2.052226619753128e-07,
+      "step": 715
+    },
+    {
+      "epoch": 0.3818666666666667,
+      "grad_norm": 0.2342621088027954,
+      "learning_rate": 3e-06,
+      "loss": 0.2106,
+      "policy/loss": 2.0455753002579513e-07,
+      "step": 716
+    },
+    {
+      "completion_length": 227.0625,
+      "epoch": 0.3824,
+      "grad_norm": 0.17591850459575653,
+      "learning_rate": 3e-06,
+      "loss": 0.0884,
+      "policy/loss": 0.16424919664859772,
+      "reward": 1.3333333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.5833333432674408,
+      "step": 717
+    },
+    {
+      "epoch": 0.38293333333333335,
+      "grad_norm": 0.21961064636707306,
+      "learning_rate": 3e-06,
+      "loss": 0.0839,
+      "policy/loss": 0.23112806864082813,
+      "step": 718
+    },
+    {
+      "epoch": 0.3834666666666667,
+      "grad_norm": 0.16486996412277222,
+      "learning_rate": 3e-06,
+      "loss": 0.0875,
+      "policy/loss": 0.1624389812350273,
+      "step": 719
+    },
+    {
+      "epoch": 0.384,
+      "grad_norm": 0.23164452612400055,
+      "learning_rate": 3e-06,
+      "loss": 0.0838,
+      "policy/loss": 0.23101993836462498,
+      "step": 720
+    },
+    {
+      "completion_length": 235.0729217529297,
+      "epoch": 0.38453333333333334,
+      "grad_norm": 0.2219366580247879,
+      "learning_rate": 3e-06,
+      "loss": -0.0758,
+      "policy/loss": -0.15801919624209404,
+      "reward": 1.1718750298023224,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.458333320915699,
+      "step": 721
+    },
+    {
+      "epoch": 0.38506666666666667,
+      "grad_norm": 0.22796571254730225,
+      "learning_rate": 3e-06,
+      "loss": -0.1282,
+      "policy/loss": -0.5740036219358444,
+      "step": 722
+    },
+    {
+      "epoch": 0.3856,
+      "grad_norm": 0.2473136931657791,
+      "learning_rate": 3e-06,
+      "loss": -0.0774,
+      "policy/loss": -0.16650189086794853,
+      "step": 723
+    },
+    {
+      "epoch": 0.38613333333333333,
+      "grad_norm": 0.18547560274600983,
+      "learning_rate": 3e-06,
+      "loss": -0.1287,
+      "policy/loss": -0.5720660537481308,
+      "step": 724
+    },
+    {
+      "completion_length": 234.8229217529297,
+      "epoch": 0.38666666666666666,
+      "grad_norm": 0.2669525146484375,
+      "learning_rate": 3e-06,
+      "loss": -0.004,
+      "policy/loss": -0.24639969691634178,
+      "reward": 1.7552083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.734375,
+      "rewards/correctness_reward_func_math": 1.0208333134651184,
+      "step": 725
+    },
+    {
+      "epoch": 0.3872,
+      "grad_norm": 0.2363949567079544,
+      "learning_rate": 3e-06,
+      "loss": -0.0917,
+      "policy/loss": 0.1728111281991005,
+      "step": 726
+    },
+    {
+      "epoch": 0.3877333333333333,
+      "grad_norm": 0.3003232777118683,
+      "learning_rate": 3e-06,
+      "loss": -0.0049,
+      "policy/loss": -0.24645743146538734,
+      "step": 727
+    },
+    {
+      "epoch": 0.38826666666666665,
+      "grad_norm": 0.2392137348651886,
+      "learning_rate": 3e-06,
+      "loss": -0.0911,
+      "policy/loss": 0.17325777560472488,
+      "step": 728
+    },
+    {
+      "completion_length": 249.53125762939453,
+      "epoch": 0.3888,
+      "grad_norm": 0.32923659682273865,
+      "learning_rate": 3e-06,
+      "loss": 0.0716,
+      "policy/loss": -0.03920064769521048,
+      "reward": 1.3854166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6666666567325592,
+      "step": 729
+    },
+    {
+      "epoch": 0.3893333333333333,
+      "grad_norm": 0.22741077840328217,
+      "learning_rate": 3e-06,
+      "loss": -0.0747,
+      "policy/loss": -0.026649763041618257,
+      "step": 730
+    },
+    {
+      "epoch": 0.38986666666666664,
+      "grad_norm": 0.3120366930961609,
+      "learning_rate": 3e-06,
+      "loss": 0.0705,
+      "policy/loss": -0.040896317960658735,
+      "step": 731
+    },
+    {
+      "epoch": 0.3904,
+      "grad_norm": 0.25224408507347107,
+      "learning_rate": 3e-06,
+      "loss": -0.0763,
+      "policy/loss": -0.02931369581659027,
+      "step": 732
+    },
+    {
+      "completion_length": 239.59375762939453,
+      "epoch": 0.39093333333333335,
+      "grad_norm": 0.22424185276031494,
+      "learning_rate": 3e-06,
+      "loss": -0.0701,
+      "policy/loss": -0.3709869310259819,
+      "reward": 1.703125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 1.0208333134651184,
+      "step": 733
+    },
+    {
+      "epoch": 0.3914666666666667,
+      "grad_norm": 0.26824042201042175,
+      "learning_rate": 3e-06,
+      "loss": 0.0086,
+      "policy/loss": -0.3389737084507942,
+      "step": 734
+    },
+    {
+      "epoch": 0.392,
+      "grad_norm": 0.21349477767944336,
+      "learning_rate": 3e-06,
+      "loss": -0.0705,
+      "policy/loss": -0.37081485986709595,
+      "step": 735
+    },
+    {
+      "epoch": 0.39253333333333335,
+      "grad_norm": 0.2605399191379547,
+      "learning_rate": 3e-06,
+      "loss": 0.0076,
+      "policy/loss": -0.3422471582889557,
+      "step": 736
+    },
+    {
+      "completion_length": 239.25,
+      "epoch": 0.3930666666666667,
+      "grad_norm": 0.23989462852478027,
+      "learning_rate": 3e-06,
+      "loss": 0.0897,
+      "policy/loss": 0.3605782234307302,
+      "reward": 1.2500000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5625,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 737
+    },
+    {
+      "epoch": 0.3936,
+      "grad_norm": 0.23979414999485016,
+      "learning_rate": 3e-06,
+      "loss": 0.0901,
+      "policy/loss": 0.43201508023697777,
+      "step": 738
+    },
+    {
+      "epoch": 0.39413333333333334,
+      "grad_norm": 0.2459549456834793,
+      "learning_rate": 3e-06,
+      "loss": 0.0885,
+      "policy/loss": 0.36010603543064246,
+      "step": 739
+    },
+    {
+      "epoch": 0.39466666666666667,
+      "grad_norm": 0.2841685712337494,
+      "learning_rate": 3e-06,
+      "loss": 0.0898,
+      "policy/loss": 0.4315406272749325,
+      "step": 740
+    },
+    {
+      "completion_length": 245.11459350585938,
+      "epoch": 0.3952,
+      "grad_norm": 0.1912776678800583,
+      "learning_rate": 3e-06,
+      "loss": -0.1455,
+      "policy/loss": 1.2355164713540034e-07,
+      "reward": 1.6666667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.9791666865348816,
+      "step": 741
+    },
+    {
+      "epoch": 0.3957333333333333,
+      "grad_norm": 0.22028124332427979,
+      "learning_rate": 3e-06,
+      "loss": -0.1117,
+      "policy/loss": 1.437507002322036e-07,
+      "step": 742
+    },
+    {
+      "epoch": 0.39626666666666666,
+      "grad_norm": 0.19148655235767365,
+      "learning_rate": 3e-06,
+      "loss": -0.1455,
+      "policy/loss": 1.2355298295574357e-07,
+      "step": 743
+    },
+    {
+      "epoch": 0.3968,
+      "grad_norm": 0.2024078369140625,
+      "learning_rate": 3e-06,
+      "loss": -0.1111,
+      "policy/loss": 1.436968055656962e-07,
+      "step": 744
+    },
+    {
+      "completion_length": 238.00000762939453,
+      "epoch": 0.3973333333333333,
+      "grad_norm": 0.21456316113471985,
+      "learning_rate": 3e-06,
+      "loss": 0.0896,
+      "policy/loss": -0.03934025704828059,
+      "reward": 1.1614583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.4791666716337204,
+      "step": 745
+    },
+    {
+      "epoch": 0.39786666666666665,
+      "grad_norm": 0.21959269046783447,
+      "learning_rate": 3e-06,
+      "loss": 0.0245,
+      "policy/loss": -0.021886145808643676,
+      "step": 746
+    },
+    {
+      "epoch": 0.3984,
+      "grad_norm": 0.20686660706996918,
+      "learning_rate": 3e-06,
+      "loss": 0.0887,
+      "policy/loss": -0.03932552359080077,
+      "step": 747
+    },
+    {
+      "epoch": 0.3989333333333333,
+      "grad_norm": 0.2034527212381363,
+      "learning_rate": 3e-06,
+      "loss": 0.0236,
+      "policy/loss": -0.029663667199294252,
+      "step": 748
+    },
+    {
+      "completion_length": 230.64584350585938,
+      "epoch": 0.3994666666666667,
+      "grad_norm": 0.3155837655067444,
+      "learning_rate": 3e-06,
+      "loss": -0.0377,
+      "policy/loss": 0.19655586939731506,
+      "reward": 1.2604166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.6041666567325592,
+      "step": 749
+    },
+    {
+      "epoch": 0.4,
+      "grad_norm": 0.2864879369735718,
+      "learning_rate": 3e-06,
+      "loss": -0.033,
+      "policy/loss": -0.4969689112416482,
+      "step": 750
+    },
+    {
+      "epoch": 0.40053333333333335,
+      "grad_norm": 0.2631821036338806,
+      "learning_rate": 3e-06,
+      "loss": -0.0379,
+      "policy/loss": 0.19907966429405022,
+      "step": 751
+    },
+    {
+      "epoch": 0.4010666666666667,
+      "grad_norm": 0.28791871666908264,
+      "learning_rate": 3e-06,
+      "loss": -0.0347,
+      "policy/loss": -0.5022347133819629,
+      "step": 752
+    },
+    {
+      "completion_length": 245.92709350585938,
+      "epoch": 0.4016,
+      "grad_norm": 0.19570551812648773,
+      "learning_rate": 3e-06,
+      "loss": 0.2501,
+      "policy/loss": 1.1511862574309362e-07,
+      "reward": 1.1666666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333134651184,
+      "rewards/correctness_reward_func_math": 0.4583333283662796,
+      "step": 753
+    },
+    {
+      "epoch": 0.40213333333333334,
+      "grad_norm": 0.18154452741146088,
+      "learning_rate": 3e-06,
+      "loss": -0.006,
+      "policy/loss": 1.135801070972775e-07,
+      "step": 754
+    },
+    {
+      "epoch": 0.4026666666666667,
+      "grad_norm": 0.19834817945957184,
+      "learning_rate": 3e-06,
+      "loss": 0.2491,
+      "policy/loss": 1.1503528618561631e-07,
+      "step": 755
+    },
+    {
+      "epoch": 0.4032,
+      "grad_norm": 0.17848576605319977,
+      "learning_rate": 3e-06,
+      "loss": -0.0068,
+      "policy/loss": 1.133577072209846e-07,
+      "step": 756
+    },
+    {
+      "completion_length": 242.67708587646484,
+      "epoch": 0.40373333333333333,
+      "grad_norm": 0.39598095417022705,
+      "learning_rate": 3e-06,
+      "loss": 0.0825,
+      "policy/loss": 0.13766389903217657,
+      "reward": 1.1875000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.4999999850988388,
+      "step": 757
+    },
+    {
+      "epoch": 0.40426666666666666,
+      "grad_norm": 0.2848465144634247,
+      "learning_rate": 3e-06,
+      "loss": -0.0002,
+      "policy/loss": 0.5759678319449542,
+      "step": 758
+    },
+    {
+      "epoch": 0.4048,
+      "grad_norm": 0.3812107741832733,
+      "learning_rate": 3e-06,
+      "loss": 0.0819,
+      "policy/loss": 0.1331039350365586,
+      "step": 759
+    },
+    {
+      "epoch": 0.4053333333333333,
+      "grad_norm": 0.26270583271980286,
+      "learning_rate": 3e-06,
+      "loss": -0.0013,
+      "policy/loss": 0.5685074762563822,
+      "step": 760
+    },
+    {
+      "completion_length": 229.7604217529297,
+      "epoch": 0.40586666666666665,
+      "grad_norm": 0.18071258068084717,
+      "learning_rate": 3e-06,
+      "loss": 0.1774,
+      "policy/loss": 9.523223098995004e-08,
+      "reward": 1.3593750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.7083333432674408,
+      "step": 761
+    },
+    {
+      "epoch": 0.4064,
+      "grad_norm": 0.19659273326396942,
+      "learning_rate": 3e-06,
+      "loss": 0.0567,
+      "policy/loss": 6.971463406557632e-08,
+      "step": 762
+    },
+    {
+      "epoch": 0.4069333333333333,
+      "grad_norm": 0.18020111322402954,
+      "learning_rate": 3e-06,
+      "loss": 0.1768,
+      "policy/loss": 9.515176913055257e-08,
+      "step": 763
+    },
+    {
+      "epoch": 0.40746666666666664,
+      "grad_norm": 0.20263515412807465,
+      "learning_rate": 3e-06,
+      "loss": 0.0558,
+      "policy/loss": 6.974877564402959e-08,
+      "step": 764
+    },
+    {
+      "completion_length": 238.05208587646484,
+      "epoch": 0.408,
+      "grad_norm": 0.19037672877311707,
+      "learning_rate": 3e-06,
+      "loss": 0.1293,
+      "policy/loss": 0.3416631903572238,
+      "reward": 1.1770833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.5208333432674408,
+      "step": 765
+    },
+    {
+      "epoch": 0.40853333333333336,
+      "grad_norm": 0.20022588968276978,
+      "learning_rate": 3e-06,
+      "loss": 0.0289,
+      "policy/loss": -0.35142956523971947,
+      "step": 766
+    },
+    {
+      "epoch": 0.4090666666666667,
+      "grad_norm": 0.19158408045768738,
+      "learning_rate": 3e-06,
+      "loss": 0.1283,
+      "policy/loss": 0.34093455335791134,
+      "step": 767
+    },
+    {
+      "epoch": 0.4096,
+      "grad_norm": 0.20001749694347382,
+      "learning_rate": 3e-06,
+      "loss": 0.0287,
+      "policy/loss": -0.35285539786557507,
+      "step": 768
+    },
+    {
+      "completion_length": 232.06250762939453,
+      "epoch": 0.41013333333333335,
+      "grad_norm": 0.25254133343696594,
+      "learning_rate": 3e-06,
+      "loss": -0.0962,
+      "policy/loss": 0.05859868137355306,
+      "reward": 1.4791666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 769
+    },
+    {
+      "epoch": 0.4106666666666667,
+      "grad_norm": 0.19745449721813202,
+      "learning_rate": 3e-06,
+      "loss": -0.0007,
+      "policy/loss": 0.3594794253203766,
+      "step": 770
+    },
+    {
+      "epoch": 0.4112,
+      "grad_norm": 0.24277363717556,
+      "learning_rate": 3e-06,
+      "loss": -0.0976,
+      "policy/loss": 0.05682444475330328,
+      "step": 771
+    },
+    {
+      "epoch": 0.41173333333333334,
+      "grad_norm": 0.19897107779979706,
+      "learning_rate": 3e-06,
+      "loss": 0.0002,
+      "policy/loss": 0.35908505116008627,
+      "step": 772
+    },
+    {
+      "completion_length": 235.5104217529297,
+      "epoch": 0.41226666666666667,
+      "grad_norm": 0.21504105627536774,
+      "learning_rate": 3e-06,
+      "loss": 0.1749,
+      "policy/loss": 0.1460293084383011,
+      "reward": 1.6197916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.875,
+      "step": 773
+    },
+    {
+      "epoch": 0.4128,
+      "grad_norm": 0.23847471177577972,
+      "learning_rate": 3e-06,
+      "loss": 0.0413,
+      "policy/loss": 0.07271292060613632,
+      "step": 774
+    },
+    {
+      "epoch": 0.41333333333333333,
+      "grad_norm": 0.20713266730308533,
+      "learning_rate": 3e-06,
+      "loss": 0.1737,
+      "policy/loss": 0.1439262479543686,
+      "step": 775
+    },
+    {
+      "epoch": 0.41386666666666666,
+      "grad_norm": 0.2304028868675232,
+      "learning_rate": 3e-06,
+      "loss": 0.0422,
+      "policy/loss": 0.0737595334649086,
+      "step": 776
+    },
+    {
+      "completion_length": 239.03125762939453,
+      "epoch": 0.4144,
+      "grad_norm": 0.26834917068481445,
+      "learning_rate": 3e-06,
+      "loss": -0.0639,
+      "policy/loss": -0.16935820035354254,
+      "reward": 1.3697916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.640625,
+      "rewards/correctness_reward_func_math": 0.7291666865348816,
+      "step": 777
+    },
+    {
+      "epoch": 0.4149333333333333,
+      "grad_norm": 0.26082441210746765,
+      "learning_rate": 3e-06,
+      "loss": -0.1582,
+      "policy/loss": -0.2044418847444618,
+      "step": 778
+    },
+    {
+      "epoch": 0.41546666666666665,
+      "grad_norm": 0.23369474709033966,
+      "learning_rate": 3e-06,
+      "loss": -0.0648,
+      "policy/loss": -0.169612995283849,
+      "step": 779
+    },
+    {
+      "epoch": 0.416,
+      "grad_norm": 0.3315592408180237,
+      "learning_rate": 3e-06,
+      "loss": -0.1582,
+      "policy/loss": -0.20628907737417634,
+      "step": 780
+    },
+    {
+      "completion_length": 234.48959350585938,
+      "epoch": 0.4165333333333333,
+      "grad_norm": 0.24018454551696777,
+      "learning_rate": 3e-06,
+      "loss": -0.0278,
+      "policy/loss": 1.3670604914750584e-07,
+      "reward": 1.0781250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6197916567325592,
+      "rewards/correctness_reward_func_math": 0.4583333432674408,
+      "step": 781
+    },
+    {
+      "epoch": 0.41706666666666664,
+      "grad_norm": 0.2783350348472595,
+      "learning_rate": 3e-06,
+      "loss": -0.1388,
+      "policy/loss": 1.3305236734595383e-07,
+      "step": 782
+    },
+    {
+      "epoch": 0.4176,
+      "grad_norm": 0.1992199569940567,
+      "learning_rate": 3e-06,
+      "loss": -0.0283,
+      "policy/loss": 1.3659067832350047e-07,
+      "step": 783
+    },
+    {
+      "epoch": 0.41813333333333336,
+      "grad_norm": 0.2868705987930298,
+      "learning_rate": 3e-06,
+      "loss": -0.14,
+      "policy/loss": 1.329715608733295e-07,
+      "step": 784
+    },
+    {
+      "completion_length": 244.30208587646484,
+      "epoch": 0.4186666666666667,
+      "grad_norm": 0.17421875894069672,
+      "learning_rate": 3e-06,
+      "loss": -0.0427,
+      "policy/loss": -0.12263695150613785,
+      "reward": 1.5833333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.8958333134651184,
+      "step": 785
+    },
+    {
+      "epoch": 0.4192,
+      "grad_norm": 0.22551336884498596,
+      "learning_rate": 3e-06,
+      "loss": 0.0595,
+      "policy/loss": -0.14861398935317993,
+      "step": 786
+    },
+    {
+      "epoch": 0.41973333333333335,
+      "grad_norm": 0.16258804500102997,
+      "learning_rate": 3e-06,
+      "loss": -0.0437,
+      "policy/loss": -0.12815168499946594,
+      "step": 787
+    },
+    {
+      "epoch": 0.4202666666666667,
+      "grad_norm": 0.18054302036762238,
+      "learning_rate": 3e-06,
+      "loss": 0.0578,
+      "policy/loss": -0.15084592998027802,
+      "step": 788
+    },
+    {
+      "completion_length": 245.31250762939453,
+      "epoch": 0.4208,
+      "grad_norm": 0.20222650468349457,
+      "learning_rate": 3e-06,
+      "loss": 0.1915,
+      "policy/loss": 0.3907765646256003,
+      "reward": 1.2604166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.5416666716337204,
+      "step": 789
+    },
+    {
+      "epoch": 0.42133333333333334,
+      "grad_norm": 0.21082231402397156,
+      "learning_rate": 3e-06,
+      "loss": -0.0307,
+      "policy/loss": 0.07652332673650619,
+      "step": 790
+    },
+    {
+      "epoch": 0.42186666666666667,
+      "grad_norm": 0.1909177005290985,
+      "learning_rate": 3e-06,
+      "loss": 0.1901,
+      "policy/loss": 0.38715000939237143,
+      "step": 791
+    },
+    {
+      "epoch": 0.4224,
+      "grad_norm": 0.28727835416793823,
+      "learning_rate": 3e-06,
+      "loss": -0.0309,
+      "policy/loss": 0.0782780875628184,
+      "step": 792
+    },
+    {
+      "completion_length": 237.0729217529297,
+      "epoch": 0.42293333333333333,
+      "grad_norm": 0.18668697774410248,
+      "learning_rate": 3e-06,
+      "loss": -0.0782,
+      "policy/loss": -0.055228887490374134,
+      "reward": 1.6354167461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.9166666567325592,
+      "step": 793
+    },
+    {
+      "epoch": 0.42346666666666666,
+      "grad_norm": 0.20354336500167847,
+      "learning_rate": 3e-06,
+      "loss": -0.0428,
+      "policy/loss": -0.05352627244170094,
+      "step": 794
+    },
+    {
+      "epoch": 0.424,
+      "grad_norm": 0.1748165786266327,
+      "learning_rate": 3e-06,
+      "loss": -0.0779,
+      "policy/loss": -0.0536237380938136,
+      "step": 795
+    },
+    {
+      "epoch": 0.4245333333333333,
+      "grad_norm": 0.20739860832691193,
+      "learning_rate": 3e-06,
+      "loss": -0.0444,
+      "policy/loss": -0.05640781430471442,
+      "step": 796
+    },
+    {
+      "completion_length": 236.52084350585938,
+      "epoch": 0.42506666666666665,
+      "grad_norm": 0.18611635267734528,
+      "learning_rate": 3e-06,
+      "loss": -0.0897,
+      "policy/loss": -0.22454567284886107,
+      "reward": 1.4166666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7291666865348816,
+      "step": 797
+    },
+    {
+      "epoch": 0.4256,
+      "grad_norm": 0.24596358835697174,
+      "learning_rate": 3e-06,
+      "loss": -0.1218,
+      "policy/loss": -0.03075892154045512,
+      "step": 798
+    },
+    {
+      "epoch": 0.4261333333333333,
+      "grad_norm": 0.1900155246257782,
+      "learning_rate": 3e-06,
+      "loss": -0.0904,
+      "policy/loss": -0.22583954072060308,
+      "step": 799
+    },
+    {
+      "epoch": 0.4266666666666667,
+      "grad_norm": 0.2416977435350418,
+      "learning_rate": 3e-06,
+      "loss": -0.1226,
+      "policy/loss": -0.03404603189652189,
+      "step": 800
+    },
+    {
+      "completion_length": 236.73958587646484,
+      "epoch": 0.4272,
+      "grad_norm": 0.2146853357553482,
+      "learning_rate": 3e-06,
+      "loss": -0.1689,
+      "policy/loss": -0.4723799116909504,
+      "reward": 1.1770833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6770833432674408,
+      "rewards/correctness_reward_func_math": 0.4999999850988388,
+      "step": 801
+    },
+    {
+      "epoch": 0.42773333333333335,
+      "grad_norm": 0.3377997577190399,
+      "learning_rate": 3e-06,
+      "loss": 0.034,
+      "policy/loss": 0.17084988206624985,
+      "step": 802
+    },
+    {
+      "epoch": 0.4282666666666667,
+      "grad_norm": 0.22043129801750183,
+      "learning_rate": 3e-06,
+      "loss": -0.1699,
+      "policy/loss": -0.47594442404806614,
+      "step": 803
+    },
+    {
+      "epoch": 0.4288,
+      "grad_norm": 0.3553898334503174,
+      "learning_rate": 3e-06,
+      "loss": 0.0331,
+      "policy/loss": 0.1661277934908867,
+      "step": 804
+    },
+    {
+      "completion_length": 236.83333587646484,
+      "epoch": 0.42933333333333334,
+      "grad_norm": 0.23048445582389832,
+      "learning_rate": 3e-06,
+      "loss": -0.2511,
+      "policy/loss": -0.4809289901044469,
+      "reward": 1.2083333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.5208333246409893,
+      "step": 805
+    },
+    {
+      "epoch": 0.4298666666666667,
+      "grad_norm": 0.2216334491968155,
+      "learning_rate": 3e-06,
+      "loss": -0.021,
+      "policy/loss": -0.0954752945516617,
+      "step": 806
+    },
+    {
+      "epoch": 0.4304,
+      "grad_norm": 0.2225360870361328,
+      "learning_rate": 3e-06,
+      "loss": -0.2519,
+      "policy/loss": -0.480033519713718,
+      "step": 807
+    },
+    {
+      "epoch": 0.43093333333333333,
+      "grad_norm": 0.2089131623506546,
+      "learning_rate": 3e-06,
+      "loss": -0.0212,
+      "policy/loss": -0.09579136309027803,
+      "step": 808
+    },
+    {
+      "completion_length": 237.625,
+      "epoch": 0.43146666666666667,
+      "grad_norm": 0.24497929215431213,
+      "learning_rate": 3e-06,
+      "loss": -0.0402,
+      "policy/loss": 0.08236581448672808,
+      "reward": 1.2135416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.5625,
+      "step": 809
+    },
+    {
+      "epoch": 0.432,
+      "grad_norm": 0.23534169793128967,
+      "learning_rate": 3e-06,
+      "loss": 0.1973,
+      "policy/loss": 0.10613824036400388,
+      "step": 810
+    },
+    {
+      "epoch": 0.4325333333333333,
+      "grad_norm": 0.25999411940574646,
+      "learning_rate": 3e-06,
+      "loss": -0.0407,
+      "policy/loss": 0.0836296267622032,
+      "step": 811
+    },
+    {
+      "epoch": 0.43306666666666666,
+      "grad_norm": 0.2434770166873932,
+      "learning_rate": 3e-06,
+      "loss": 0.1974,
+      "policy/loss": 0.10595099988191947,
+      "step": 812
+    },
+    {
+      "completion_length": 242.25000762939453,
+      "epoch": 0.4336,
+      "grad_norm": 0.22790154814720154,
+      "learning_rate": 3e-06,
+      "loss": 0.0777,
+      "policy/loss": -0.051092149103091344,
+      "reward": 1.9895833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 1.2708333134651184,
+      "step": 813
+    },
+    {
+      "epoch": 0.4341333333333333,
+      "grad_norm": 0.24125166237354279,
+      "learning_rate": 3e-06,
+      "loss": -0.0558,
+      "policy/loss": 0.12950750867692307,
+      "step": 814
+    },
+    {
+      "epoch": 0.43466666666666665,
+      "grad_norm": 0.23120734095573425,
+      "learning_rate": 3e-06,
+      "loss": 0.0766,
+      "policy/loss": -0.051478823027167664,
+      "step": 815
+    },
+    {
+      "epoch": 0.4352,
+      "grad_norm": 0.23089921474456787,
+      "learning_rate": 3e-06,
+      "loss": -0.0569,
+      "policy/loss": 0.12859965541061058,
+      "step": 816
+    },
+    {
+      "completion_length": 235.06250762939453,
+      "epoch": 0.4357333333333333,
+      "grad_norm": 0.20224472880363464,
+      "learning_rate": 3e-06,
+      "loss": 0.0212,
+      "policy/loss": 0.249686848372221,
+      "reward": 1.3645833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.7083333432674408,
+      "step": 817
+    },
+    {
+      "epoch": 0.4362666666666667,
+      "grad_norm": 0.23899269104003906,
+      "learning_rate": 3e-06,
+      "loss": 0.0103,
+      "policy/loss": -0.014637991786003113,
+      "step": 818
+    },
+    {
+      "epoch": 0.4368,
+      "grad_norm": 0.20899300277233124,
+      "learning_rate": 3e-06,
+      "loss": 0.0215,
+      "policy/loss": 0.25083376094698906,
+      "step": 819
+    },
+    {
+      "epoch": 0.43733333333333335,
+      "grad_norm": 0.38294684886932373,
+      "learning_rate": 3e-06,
+      "loss": 0.0086,
+      "policy/loss": -0.02033422887325287,
+      "step": 820
+    },
+    {
+      "completion_length": 242.30209350585938,
+      "epoch": 0.4378666666666667,
+      "grad_norm": 0.18843954801559448,
+      "learning_rate": 3e-06,
+      "loss": -0.053,
+      "policy/loss": 0.04754011332988739,
+      "reward": 1.3333333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 821
+    },
+    {
+      "epoch": 0.4384,
+      "grad_norm": 0.22725170850753784,
+      "learning_rate": 3e-06,
+      "loss": 0.0237,
+      "policy/loss": 0.14425347745418549,
+      "step": 822
+    },
+    {
+      "epoch": 0.43893333333333334,
+      "grad_norm": 0.18060344457626343,
+      "learning_rate": 3e-06,
+      "loss": -0.0533,
+      "policy/loss": 0.04722793400287628,
+      "step": 823
+    },
+    {
+      "epoch": 0.43946666666666667,
+      "grad_norm": 0.19244208931922913,
+      "learning_rate": 3e-06,
+      "loss": 0.0222,
+      "policy/loss": 0.13780206441879272,
+      "step": 824
+    },
+    {
+      "completion_length": 241.8854217529297,
+      "epoch": 0.44,
+      "grad_norm": 0.19556429982185364,
+      "learning_rate": 3e-06,
+      "loss": 0.0439,
+      "policy/loss": 6.598702029236847e-08,
+      "reward": 1.0833333432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5625,
+      "rewards/correctness_reward_func_math": 0.5208333283662796,
+      "step": 825
+    },
+    {
+      "epoch": 0.44053333333333333,
+      "grad_norm": 0.18455401062965393,
+      "learning_rate": 3e-06,
+      "loss": 0.0298,
+      "policy/loss": 7.164096160749978e-08,
+      "step": 826
+    },
+    {
+      "epoch": 0.44106666666666666,
+      "grad_norm": 0.19296744465827942,
+      "learning_rate": 3e-06,
+      "loss": 0.0431,
+      "policy/loss": 6.601371360659414e-08,
+      "step": 827
+    },
+    {
+      "epoch": 0.4416,
+      "grad_norm": 0.18601466715335846,
+      "learning_rate": 3e-06,
+      "loss": 0.029,
+      "policy/loss": 7.161944104439044e-08,
+      "step": 828
+    },
+    {
+      "completion_length": 228.93750762939453,
+      "epoch": 0.4421333333333333,
+      "grad_norm": 0.23028193414211273,
+      "learning_rate": 3e-06,
+      "loss": -0.0453,
+      "policy/loss": -0.36722692039152616,
+      "reward": 1.3854167461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.7291666865348816,
+      "step": 829
+    },
+    {
+      "epoch": 0.44266666666666665,
+      "grad_norm": 0.23223842680454254,
+      "learning_rate": 3e-06,
+      "loss": 0.0644,
+      "policy/loss": 0.45156792359886166,
+      "step": 830
+    },
+    {
+      "epoch": 0.4432,
+      "grad_norm": 0.23574312031269073,
+      "learning_rate": 3e-06,
+      "loss": -0.0463,
+      "policy/loss": -0.3726921596854478,
+      "step": 831
+    },
+    {
+      "epoch": 0.4437333333333333,
+      "grad_norm": 0.2412298172712326,
+      "learning_rate": 3e-06,
+      "loss": 0.063,
+      "policy/loss": 0.44809860544972313,
+      "step": 832
+    },
+    {
+      "completion_length": 243.6354217529297,
+      "epoch": 0.44426666666666664,
+      "grad_norm": 0.27043887972831726,
+      "learning_rate": 3e-06,
+      "loss": -0.0516,
+      "policy/loss": 0.11444329808000475,
+      "reward": 1.5260416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 833
+    },
+    {
+      "epoch": 0.4448,
+      "grad_norm": 0.19070854783058167,
+      "learning_rate": 3e-06,
+      "loss": 0.0117,
+      "policy/loss": 0.13730011880397797,
+      "step": 834
+    },
+    {
+      "epoch": 0.44533333333333336,
+      "grad_norm": 0.27435094118118286,
+      "learning_rate": 3e-06,
+      "loss": -0.0523,
+      "policy/loss": 0.11219595489092171,
+      "step": 835
+    },
+    {
+      "epoch": 0.4458666666666667,
+      "grad_norm": 0.18484000861644745,
+      "learning_rate": 3e-06,
+      "loss": 0.0112,
+      "policy/loss": 0.13769777119159698,
+      "step": 836
+    },
+    {
+      "completion_length": 245.4166717529297,
+      "epoch": 0.4464,
+      "grad_norm": 0.19915015995502472,
+      "learning_rate": 3e-06,
+      "loss": -0.0324,
+      "policy/loss": 0.12906844308432142,
+      "reward": 1.5260417461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.734375,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 837
+    },
+    {
+      "epoch": 0.44693333333333335,
+      "grad_norm": 0.1790773868560791,
+      "learning_rate": 3e-06,
+      "loss": 0.067,
+      "policy/loss": 0.1735716969707113,
+      "step": 838
+    },
+    {
+      "epoch": 0.4474666666666667,
+      "grad_norm": 0.19913408160209656,
+      "learning_rate": 3e-06,
+      "loss": -0.0331,
+      "policy/loss": 0.12918221342973624,
+      "step": 839
+    },
+    {
+      "epoch": 0.448,
+      "grad_norm": 0.18741825222969055,
+      "learning_rate": 3e-06,
+      "loss": 0.0658,
+      "policy/loss": 0.1742113293422456,
+      "step": 840
+    },
+    {
+      "completion_length": 245.70834350585938,
+      "epoch": 0.44853333333333334,
+      "grad_norm": 0.22496344149112701,
+      "learning_rate": 3e-06,
+      "loss": 0.149,
+      "policy/loss": 0.13198892027139664,
+      "reward": 1.3020833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.6458333283662796,
+      "step": 841
+    },
+    {
+      "epoch": 0.44906666666666667,
+      "grad_norm": 0.21639969944953918,
+      "learning_rate": 3e-06,
+      "loss": -0.1055,
+      "policy/loss": 0.20976761355996132,
+      "step": 842
+    },
+    {
+      "epoch": 0.4496,
+      "grad_norm": 0.21312101185321808,
+      "learning_rate": 3e-06,
+      "loss": 0.1479,
+      "policy/loss": 0.13038266450166702,
+      "step": 843
+    },
+    {
+      "epoch": 0.45013333333333333,
+      "grad_norm": 0.2092207968235016,
+      "learning_rate": 3e-06,
+      "loss": -0.1073,
+      "policy/loss": 0.20650320127606392,
+      "step": 844
+    },
+    {
+      "completion_length": 242.46875762939453,
+      "epoch": 0.45066666666666666,
+      "grad_norm": 0.28899192810058594,
+      "learning_rate": 3e-06,
+      "loss": -0.0671,
+      "policy/loss": -0.17872187684140428,
+      "reward": 1.28125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 845
+    },
+    {
+      "epoch": 0.4512,
+      "grad_norm": 0.2133583277463913,
+      "learning_rate": 3e-06,
+      "loss": -0.0281,
+      "policy/loss": -0.3428852589983009,
+      "step": 846
+    },
+    {
+      "epoch": 0.4517333333333333,
+      "grad_norm": 0.250858336687088,
+      "learning_rate": 3e-06,
+      "loss": -0.0691,
+      "policy/loss": -0.17957014036985797,
+      "step": 847
+    },
+    {
+      "epoch": 0.45226666666666665,
+      "grad_norm": 0.2124241292476654,
+      "learning_rate": 3e-06,
+      "loss": -0.0296,
+      "policy/loss": -0.34453207573152156,
+      "step": 848
+    },
+    {
+      "completion_length": 239.8229217529297,
+      "epoch": 0.4528,
+      "grad_norm": 0.31598547101020813,
+      "learning_rate": 3e-06,
+      "loss": 0.0407,
+      "policy/loss": -0.07518938757309002,
+      "reward": 1.6614583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.9166666567325592,
+      "step": 849
+    },
+    {
+      "epoch": 0.4533333333333333,
+      "grad_norm": 0.26564785838127136,
+      "learning_rate": 3e-06,
+      "loss": -0.0318,
+      "policy/loss": 0.12326357741061855,
+      "step": 850
+    },
+    {
+      "epoch": 0.45386666666666664,
+      "grad_norm": 0.23161913454532623,
+      "learning_rate": 3e-06,
+      "loss": 0.0389,
+      "policy/loss": -0.07311153230685008,
+      "step": 851
+    },
+    {
+      "epoch": 0.4544,
+      "grad_norm": 0.2822839915752411,
+      "learning_rate": 3e-06,
+      "loss": -0.0344,
+      "policy/loss": 0.12474443986564054,
+      "step": 852
+    },
+    {
+      "completion_length": 233.21875,
+      "epoch": 0.45493333333333336,
+      "grad_norm": 0.4213089942932129,
+      "learning_rate": 3e-06,
+      "loss": -0.108,
+      "policy/loss": 0.2183364963653318,
+      "reward": 1.765625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 1.0833333432674408,
+      "step": 853
+    },
+    {
+      "epoch": 0.4554666666666667,
+      "grad_norm": 0.2609274983406067,
+      "learning_rate": 3e-06,
+      "loss": 0.0353,
+      "policy/loss": -0.26877832582792394,
+      "step": 854
+    },
+    {
+      "epoch": 0.456,
+      "grad_norm": 0.44496291875839233,
+      "learning_rate": 3e-06,
+      "loss": -0.1087,
+      "policy/loss": 0.2168973124980127,
+      "step": 855
+    },
+    {
+      "epoch": 0.45653333333333335,
+      "grad_norm": 0.25551292300224304,
+      "learning_rate": 3e-06,
+      "loss": 0.0346,
+      "policy/loss": -0.2686180192208809,
+      "step": 856
+    },
+    {
+      "completion_length": 245.09375762939453,
+      "epoch": 0.4570666666666667,
+      "grad_norm": 0.16654187440872192,
+      "learning_rate": 3e-06,
+      "loss": -0.0174,
+      "policy/loss": -0.07446612952547937,
+      "reward": 1.4166667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.7916666865348816,
+      "step": 857
+    },
+    {
+      "epoch": 0.4576,
+      "grad_norm": 0.21875129640102386,
+      "learning_rate": 3e-06,
+      "loss": -0.0223,
+      "policy/loss": 0.12400188110427024,
+      "step": 858
+    },
+    {
+      "epoch": 0.45813333333333334,
+      "grad_norm": 0.18178246915340424,
+      "learning_rate": 3e-06,
+      "loss": -0.0173,
+      "policy/loss": -0.07515195540607778,
+      "step": 859
+    },
+    {
+      "epoch": 0.45866666666666667,
+      "grad_norm": 0.2251964807510376,
+      "learning_rate": 3e-06,
+      "loss": -0.0229,
+      "policy/loss": 0.12014182465526346,
+      "step": 860
+    },
+    {
+      "completion_length": 236.5729217529297,
+      "epoch": 0.4592,
+      "grad_norm": 0.2229735255241394,
+      "learning_rate": 3e-06,
+      "loss": -0.2106,
+      "policy/loss": 1.5061872460364611e-07,
+      "reward": 1.5104166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7395833432674408,
+      "rewards/correctness_reward_func_math": 0.7708333134651184,
+      "step": 861
+    },
+    {
+      "epoch": 0.4597333333333333,
+      "grad_norm": 0.21258816123008728,
+      "learning_rate": 3e-06,
+      "loss": -0.0399,
+      "policy/loss": 1.7217297454408254e-07,
+      "step": 862
+    },
+    {
+      "epoch": 0.46026666666666666,
+      "grad_norm": 0.2205979824066162,
+      "learning_rate": 3e-06,
+      "loss": -0.212,
+      "policy/loss": 1.506366089643052e-07,
+      "step": 863
+    },
+    {
+      "epoch": 0.4608,
+      "grad_norm": 0.21846571564674377,
+      "learning_rate": 3e-06,
+      "loss": -0.0406,
+      "policy/loss": 1.723254996477408e-07,
+      "step": 864
+    },
+    {
+      "completion_length": 231.58334350585938,
+      "epoch": 0.4613333333333333,
+      "grad_norm": 0.40573060512542725,
+      "learning_rate": 3e-06,
+      "loss": -0.0042,
+      "policy/loss": 6.417847231432461e-08,
+      "reward": 1.0937500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.59375,
+      "rewards/correctness_reward_func_math": 0.5000000149011612,
+      "step": 865
+    },
+    {
+      "epoch": 0.46186666666666665,
+      "grad_norm": 0.1811976432800293,
+      "learning_rate": 3e-06,
+      "loss": -0.2286,
+      "policy/loss": 6.106618499757133e-08,
+      "step": 866
+    },
+    {
+      "epoch": 0.4624,
+      "grad_norm": 0.4072909355163574,
+      "learning_rate": 3e-06,
+      "loss": -0.0039,
+      "policy/loss": 6.418269293817502e-08,
+      "step": 867
+    },
+    {
+      "epoch": 0.4629333333333333,
+      "grad_norm": 0.17734241485595703,
+      "learning_rate": 3e-06,
+      "loss": -0.2288,
+      "policy/loss": 6.119351780853322e-08,
+      "step": 868
+    },
+    {
+      "completion_length": 241.21875762939453,
+      "epoch": 0.4634666666666667,
+      "grad_norm": 0.25296226143836975,
+      "learning_rate": 3e-06,
+      "loss": -0.1126,
+      "policy/loss": -0.10789043456313685,
+      "reward": 1.4322916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 869
+    },
+    {
+      "epoch": 0.464,
+      "grad_norm": 0.22122249007225037,
+      "learning_rate": 3e-06,
+      "loss": -0.1826,
+      "policy/loss": -0.07928023016521557,
+      "step": 870
+    },
+    {
+      "epoch": 0.46453333333333335,
+      "grad_norm": 0.24557143449783325,
+      "learning_rate": 3e-06,
+      "loss": -0.1135,
+      "policy/loss": -0.10891830925023704,
+      "step": 871
+    },
+    {
+      "epoch": 0.4650666666666667,
+      "grad_norm": 0.23328159749507904,
+      "learning_rate": 3e-06,
+      "loss": -0.1847,
+      "policy/loss": -0.0799646330227759,
+      "step": 872
+    },
+    {
+      "completion_length": 246.6354217529297,
+      "epoch": 0.4656,
+      "grad_norm": 0.22445349395275116,
+      "learning_rate": 3e-06,
+      "loss": 0.1032,
+      "policy/loss": 0.046091265976428986,
+      "reward": 1.1614583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6197916567325592,
+      "rewards/correctness_reward_func_math": 0.5416666716337204,
+      "step": 873
+    },
+    {
+      "epoch": 0.46613333333333334,
+      "grad_norm": 0.2462788075208664,
+      "learning_rate": 3e-06,
+      "loss": -0.0128,
+      "policy/loss": -0.1035822182893753,
+      "step": 874
+    },
+    {
+      "epoch": 0.4666666666666667,
+      "grad_norm": 0.23311951756477356,
+      "learning_rate": 3e-06,
+      "loss": 0.1026,
+      "policy/loss": 0.03655530512332916,
+      "step": 875
+    },
+    {
+      "epoch": 0.4672,
+      "grad_norm": 0.24600960314273834,
+      "learning_rate": 3e-06,
+      "loss": -0.014,
+      "policy/loss": -0.101731451228261,
+      "step": 876
+    },
+    {
+      "completion_length": 228.7604217529297,
+      "epoch": 0.46773333333333333,
+      "grad_norm": 0.20438914000988007,
+      "learning_rate": 3e-06,
+      "loss": -0.0288,
+      "policy/loss": -0.2873252842357559,
+      "reward": 1.6979167461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.9791666865348816,
+      "step": 877
+    },
+    {
+      "epoch": 0.46826666666666666,
+      "grad_norm": 0.23764899373054504,
+      "learning_rate": 3e-06,
+      "loss": -0.0691,
+      "policy/loss": 0.2242412884933671,
+      "step": 878
+    },
+    {
+      "epoch": 0.4688,
+      "grad_norm": 0.1867373138666153,
+      "learning_rate": 3e-06,
+      "loss": -0.0296,
+      "policy/loss": -0.28851266828346667,
+      "step": 879
+    },
+    {
+      "epoch": 0.4693333333333333,
+      "grad_norm": 0.3221866190433502,
+      "learning_rate": 3e-06,
+      "loss": -0.0699,
+      "policy/loss": 0.2233772149562938,
+      "step": 880
+    },
+    {
+      "completion_length": 245.17708587646484,
+      "epoch": 0.46986666666666665,
+      "grad_norm": 0.20719191431999207,
+      "learning_rate": 3e-06,
+      "loss": 0.1246,
+      "policy/loss": -0.06787446472188208,
+      "reward": 1.28125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 881
+    },
+    {
+      "epoch": 0.4704,
+      "grad_norm": 0.251486599445343,
+      "learning_rate": 3e-06,
+      "loss": 0.0363,
+      "policy/loss": -0.20987871469493768,
+      "step": 882
+    },
+    {
+      "epoch": 0.4709333333333333,
+      "grad_norm": 0.22274097800254822,
+      "learning_rate": 3e-06,
+      "loss": 0.1244,
+      "policy/loss": -0.06776041127945831,
+      "step": 883
+    },
+    {
+      "epoch": 0.47146666666666665,
+      "grad_norm": 0.24799731373786926,
+      "learning_rate": 3e-06,
+      "loss": 0.0345,
+      "policy/loss": -0.21052149120901475,
+      "step": 884
+    },
+    {
+      "completion_length": 246.46875762939453,
+      "epoch": 0.472,
+      "grad_norm": 0.22924676537513733,
+      "learning_rate": 3e-06,
+      "loss": 0.01,
+      "policy/loss": -0.23768853768706322,
+      "reward": 1.4166666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.6666666567325592,
+      "step": 885
+    },
+    {
+      "epoch": 0.47253333333333336,
+      "grad_norm": 0.2185690551996231,
+      "learning_rate": 3e-06,
+      "loss": -0.0941,
+      "policy/loss": -0.2054544985294342,
+      "step": 886
+    },
+    {
+      "epoch": 0.4730666666666667,
+      "grad_norm": 0.2893002927303314,
+      "learning_rate": 3e-06,
+      "loss": 0.0096,
+      "policy/loss": -0.2371270190924406,
+      "step": 887
+    },
+    {
+      "epoch": 0.4736,
+      "grad_norm": 0.2189503312110901,
+      "learning_rate": 3e-06,
+      "loss": -0.095,
+      "policy/loss": -0.2050524652004242,
+      "step": 888
+    },
+    {
+      "completion_length": 235.3229217529297,
+      "epoch": 0.47413333333333335,
+      "grad_norm": 0.2725105881690979,
+      "learning_rate": 3e-06,
+      "loss": 0.0575,
+      "policy/loss": -0.16362931503570444,
+      "reward": 1.5312500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 889
+    },
+    {
+      "epoch": 0.4746666666666667,
+      "grad_norm": 0.3063274621963501,
+      "learning_rate": 3e-06,
+      "loss": -0.1597,
+      "policy/loss": -0.06129905398989166,
+      "step": 890
+    },
+    {
+      "epoch": 0.4752,
+      "grad_norm": 0.27613145112991333,
+      "learning_rate": 3e-06,
+      "loss": 0.057,
+      "policy/loss": -0.16092488842478758,
+      "step": 891
+    },
+    {
+      "epoch": 0.47573333333333334,
+      "grad_norm": 0.3090425431728363,
+      "learning_rate": 3e-06,
+      "loss": -0.1595,
+      "policy/loss": -0.06160818971732951,
+      "step": 892
+    },
+    {
+      "completion_length": 243.30209350585938,
+      "epoch": 0.47626666666666667,
+      "grad_norm": 0.2546744644641876,
+      "learning_rate": 3e-06,
+      "loss": -0.0992,
+      "policy/loss": -0.3454781323671341,
+      "reward": 1.5520833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.833333358168602,
+      "step": 893
+    },
+    {
+      "epoch": 0.4768,
+      "grad_norm": 0.25699883699417114,
+      "learning_rate": 3e-06,
+      "loss": -0.0585,
+      "policy/loss": -0.04785306751728058,
+      "step": 894
+    },
+    {
+      "epoch": 0.47733333333333333,
+      "grad_norm": 0.2282678633928299,
+      "learning_rate": 3e-06,
+      "loss": -0.0996,
+      "policy/loss": -0.3443804383277893,
+      "step": 895
+    },
+    {
+      "epoch": 0.47786666666666666,
+      "grad_norm": 0.2575819790363312,
+      "learning_rate": 3e-06,
+      "loss": -0.0591,
+      "policy/loss": -0.04850979149341583,
+      "step": 896
+    },
+    {
+      "completion_length": 243.0416717529297,
+      "epoch": 0.4784,
+      "grad_norm": 0.1871996968984604,
+      "learning_rate": 3e-06,
+      "loss": -0.1593,
+      "policy/loss": -0.09475562277714289,
+      "reward": 1.3854166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 897
+    },
+    {
+      "epoch": 0.4789333333333333,
+      "grad_norm": 0.24739812314510345,
+      "learning_rate": 3e-06,
+      "loss": -0.0852,
+      "policy/loss": 0.281947404489685,
+      "step": 898
+    },
+    {
+      "epoch": 0.47946666666666665,
+      "grad_norm": 0.18379341065883636,
+      "learning_rate": 3e-06,
+      "loss": -0.1598,
+      "policy/loss": -0.0945520133450728,
+      "step": 899
+    },
+    {
+      "epoch": 0.48,
+      "grad_norm": 0.24790869653224945,
+      "learning_rate": 3e-06,
+      "loss": -0.0858,
+      "policy/loss": 0.2815557424064252,
+      "step": 900
+    },
+    {
+      "completion_length": 244.0729217529297,
+      "epoch": 0.4805333333333333,
+      "grad_norm": 0.21094655990600586,
+      "learning_rate": 3e-06,
+      "loss": -0.0365,
+      "policy/loss": -0.13896291429209384,
+      "reward": 1.2291667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.5416666865348816,
+      "step": 901
+    },
+    {
+      "epoch": 0.48106666666666664,
+      "grad_norm": 0.15076275169849396,
+      "learning_rate": 3e-06,
+      "loss": -0.1554,
+      "policy/loss": -0.14408075151348498,
+      "step": 902
+    },
+    {
+      "epoch": 0.4816,
+      "grad_norm": 0.1861051768064499,
+      "learning_rate": 3e-06,
+      "loss": -0.0368,
+      "policy/loss": -0.13974409281741274,
+      "step": 903
+    },
+    {
+      "epoch": 0.48213333333333336,
+      "grad_norm": 0.15389135479927063,
+      "learning_rate": 3e-06,
+      "loss": -0.1551,
+      "policy/loss": -0.14381670303219352,
+      "step": 904
+    },
+    {
+      "completion_length": 239.6354217529297,
+      "epoch": 0.4826666666666667,
+      "grad_norm": 0.2858022451400757,
+      "learning_rate": 3e-06,
+      "loss": -0.0988,
+      "policy/loss": 0.13020579494612505,
+      "reward": 1.9010416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 1.1874999701976776,
+      "step": 905
+    },
+    {
+      "epoch": 0.4832,
+      "grad_norm": 0.25236976146698,
+      "learning_rate": 3e-06,
+      "loss": -0.027,
+      "policy/loss": -0.005969427497731772,
+      "step": 906
+    },
+    {
+      "epoch": 0.48373333333333335,
+      "grad_norm": 0.2874225378036499,
+      "learning_rate": 3e-06,
+      "loss": -0.1004,
+      "policy/loss": 0.12875317006336573,
+      "step": 907
+    },
+    {
+      "epoch": 0.4842666666666667,
+      "grad_norm": 0.253980815410614,
+      "learning_rate": 3e-06,
+      "loss": -0.0288,
+      "policy/loss": -0.008414373843688594,
+      "step": 908
+    },
+    {
+      "completion_length": 237.0729217529297,
+      "epoch": 0.4848,
+      "grad_norm": 0.22561638057231903,
+      "learning_rate": 3e-06,
+      "loss": -0.0047,
+      "policy/loss": -0.17401172114516683,
+      "reward": 1.4166666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.6666666641831398,
+      "step": 909
+    },
+    {
+      "epoch": 0.48533333333333334,
+      "grad_norm": 0.15801134705543518,
+      "learning_rate": 3e-06,
+      "loss": 0.0033,
+      "policy/loss": 0.053775763714075,
+      "step": 910
+    },
+    {
+      "epoch": 0.48586666666666667,
+      "grad_norm": 0.19080576300621033,
+      "learning_rate": 3e-06,
+      "loss": -0.005,
+      "policy/loss": -0.17284358919922127,
+      "step": 911
+    },
+    {
+      "epoch": 0.4864,
+      "grad_norm": 0.16059383749961853,
+      "learning_rate": 3e-06,
+      "loss": 0.0025,
+      "policy/loss": 0.048941313047558666,
+      "step": 912
+    },
+    {
+      "completion_length": 241.58333587646484,
+      "epoch": 0.48693333333333333,
+      "grad_norm": 0.2945229709148407,
+      "learning_rate": 3e-06,
+      "loss": 0.0889,
+      "policy/loss": 1.0302048281118914e-07,
+      "reward": 1.2500000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.6250000149011612,
+      "step": 913
+    },
+    {
+      "epoch": 0.48746666666666666,
+      "grad_norm": 0.21344344317913055,
+      "learning_rate": 3e-06,
+      "loss": -0.0363,
+      "policy/loss": 1.1272447864030255e-07,
+      "step": 914
+    },
+    {
+      "epoch": 0.488,
+      "grad_norm": 0.3291873335838318,
+      "learning_rate": 3e-06,
+      "loss": 0.0875,
+      "policy/loss": 1.0302290931463176e-07,
+      "step": 915
+    },
+    {
+      "epoch": 0.4885333333333333,
+      "grad_norm": 0.20195449888706207,
+      "learning_rate": 3e-06,
+      "loss": -0.0362,
+      "policy/loss": 1.1276435429863341e-07,
+      "step": 916
+    },
+    {
+      "completion_length": 232.55209350585938,
+      "epoch": 0.48906666666666665,
+      "grad_norm": 0.2992711365222931,
+      "learning_rate": 3e-06,
+      "loss": 0.0044,
+      "policy/loss": -0.24874892318621278,
+      "reward": 1.3437500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.59375,
+      "rewards/correctness_reward_func_math": 0.7499999850988388,
+      "step": 917
+    },
+    {
+      "epoch": 0.4896,
+      "grad_norm": 0.2577500641345978,
+      "learning_rate": 3e-06,
+      "loss": -0.1199,
+      "policy/loss": -0.4409802006557584,
+      "step": 918
+    },
+    {
+      "epoch": 0.4901333333333333,
+      "grad_norm": 0.3140072822570801,
+      "learning_rate": 3e-06,
+      "loss": 0.0045,
+      "policy/loss": -0.2529720766469836,
+      "step": 919
+    },
+    {
+      "epoch": 0.49066666666666664,
+      "grad_norm": 0.26412999629974365,
+      "learning_rate": 3e-06,
+      "loss": -0.1204,
+      "policy/loss": -0.44057100359350443,
+      "step": 920
+    },
+    {
+      "completion_length": 236.5,
+      "epoch": 0.4912,
+      "grad_norm": 0.28713083267211914,
+      "learning_rate": 3e-06,
+      "loss": 0.0431,
+      "policy/loss": 0.1517539881169796,
+      "reward": 1.614583432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.895833358168602,
+      "step": 921
+    },
+    {
+      "epoch": 0.49173333333333336,
+      "grad_norm": 0.212864950299263,
+      "learning_rate": 3e-06,
+      "loss": -0.1345,
+      "policy/loss": -0.03617131232749671,
+      "step": 922
+    },
+    {
+      "epoch": 0.4922666666666667,
+      "grad_norm": 0.2869960367679596,
+      "learning_rate": 3e-06,
+      "loss": 0.0421,
+      "policy/loss": 0.14661104045808315,
+      "step": 923
+    },
+    {
+      "epoch": 0.4928,
+      "grad_norm": 0.20649465918540955,
+      "learning_rate": 3e-06,
+      "loss": -0.1357,
+      "policy/loss": -0.036389636603416875,
+      "step": 924
+    },
+    {
+      "completion_length": 239.1354217529297,
+      "epoch": 0.49333333333333335,
+      "grad_norm": 0.22972430288791656,
+      "learning_rate": 3e-06,
+      "loss": -0.0974,
+      "policy/loss": -0.07963933059126305,
+      "reward": 1.7083333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 1.0208333432674408,
+      "step": 925
+    },
+    {
+      "epoch": 0.4938666666666667,
+      "grad_norm": 0.17050574719905853,
+      "learning_rate": 3e-06,
+      "loss": -0.0236,
+      "policy/loss": -0.12001092299592386,
+      "step": 926
+    },
+    {
+      "epoch": 0.4944,
+      "grad_norm": 0.21655675768852234,
+      "learning_rate": 3e-06,
+      "loss": -0.0994,
+      "policy/loss": -0.07911284281411746,
+      "step": 927
+    },
+    {
+      "epoch": 0.49493333333333334,
+      "grad_norm": 0.16148285567760468,
+      "learning_rate": 3e-06,
+      "loss": -0.0243,
+      "policy/loss": -0.12327350246847857,
+      "step": 928
+    },
+    {
+      "completion_length": 235.80208587646484,
+      "epoch": 0.49546666666666667,
+      "grad_norm": 0.35784491896629333,
+      "learning_rate": 3e-06,
+      "loss": 0.0342,
+      "policy/loss": 0.2531117808204577,
+      "reward": 1.2083333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.5208333283662796,
+      "step": 929
+    },
+    {
+      "epoch": 0.496,
+      "grad_norm": 0.2335757166147232,
+      "learning_rate": 3e-06,
+      "loss": 0.0949,
+      "policy/loss": -0.2676969398207305,
+      "step": 930
+    },
+    {
+      "epoch": 0.4965333333333333,
+      "grad_norm": 0.29578885436058044,
+      "learning_rate": 3e-06,
+      "loss": 0.0339,
+      "policy/loss": 0.2528260361257857,
+      "step": 931
+    },
+    {
+      "epoch": 0.49706666666666666,
+      "grad_norm": 0.2676309645175934,
+      "learning_rate": 3e-06,
+      "loss": 0.0946,
+      "policy/loss": -0.2669951547446061,
+      "step": 932
+    },
+    {
+      "completion_length": 245.02084350585938,
+      "epoch": 0.4976,
+      "grad_norm": 0.24281108379364014,
+      "learning_rate": 3e-06,
+      "loss": 0.0422,
+      "policy/loss": 0.39766572415828705,
+      "reward": 1.3125000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 933
+    },
+    {
+      "epoch": 0.4981333333333333,
+      "grad_norm": 0.20081695914268494,
+      "learning_rate": 3e-06,
+      "loss": -0.0722,
+      "policy/loss": -0.3149230629205704,
+      "step": 934
+    },
+    {
+      "epoch": 0.49866666666666665,
+      "grad_norm": 0.2319290190935135,
+      "learning_rate": 3e-06,
+      "loss": 0.041,
+      "policy/loss": 0.395385280251503,
+      "step": 935
+    },
+    {
+      "epoch": 0.4992,
+      "grad_norm": 0.21086059510707855,
+      "learning_rate": 3e-06,
+      "loss": -0.0726,
+      "policy/loss": -0.3136213570833206,
+      "step": 936
+    },
+    {
+      "completion_length": 241.1041717529297,
+      "epoch": 0.4997333333333333,
+      "grad_norm": 0.3151424825191498,
+      "learning_rate": 3e-06,
+      "loss": 0.0267,
+      "policy/loss": 1.303409824515711e-07,
+      "reward": 1.5833333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333134651184,
+      "rewards/correctness_reward_func_math": 0.875,
+      "step": 937
+    },
+    {
+      "epoch": 0.5002666666666666,
+      "grad_norm": 0.27698227763175964,
+      "learning_rate": 3e-06,
+      "loss": -0.084,
+      "policy/loss": 1.7112617456405133e-07,
+      "step": 938
+    },
+    {
+      "epoch": 0.5008,
+      "grad_norm": 0.29808253049850464,
+      "learning_rate": 3e-06,
+      "loss": 0.0261,
+      "policy/loss": 1.305410108898286e-07,
+      "step": 939
+    },
+    {
+      "epoch": 0.5013333333333333,
+      "grad_norm": 0.28395938873291016,
+      "learning_rate": 3e-06,
+      "loss": -0.0849,
+      "policy/loss": 1.7135144858571039e-07,
+      "step": 940
+    },
+    {
+      "completion_length": 241.0416717529297,
+      "epoch": 0.5018666666666667,
+      "grad_norm": 0.24956896901130676,
+      "learning_rate": 3e-06,
+      "loss": 0.0743,
+      "policy/loss": 0.23458636695393764,
+      "reward": 1.8854167461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 1.1666666567325592,
+      "step": 941
+    },
+    {
+      "epoch": 0.5024,
+      "grad_norm": 0.665886640548706,
+      "learning_rate": 3e-06,
+      "loss": -0.0735,
+      "policy/loss": 0.266606245791813,
+      "step": 942
+    },
+    {
+      "epoch": 0.5029333333333333,
+      "grad_norm": 0.24888962507247925,
+      "learning_rate": 3e-06,
+      "loss": 0.0741,
+      "policy/loss": 0.23349326249523372,
+      "step": 943
+    },
+    {
+      "epoch": 0.5034666666666666,
+      "grad_norm": 0.27021917700767517,
+      "learning_rate": 3e-06,
+      "loss": -0.0741,
+      "policy/loss": 0.26853412811285793,
+      "step": 944
+    },
+    {
+      "completion_length": 242.80208587646484,
+      "epoch": 0.504,
+      "grad_norm": 0.2360142022371292,
+      "learning_rate": 3e-06,
+      "loss": -0.0122,
+      "policy/loss": 0.059966754889545726,
+      "reward": 1.4062500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 945
+    },
+    {
+      "epoch": 0.5045333333333333,
+      "grad_norm": 0.2123853713274002,
+      "learning_rate": 3e-06,
+      "loss": 0.1175,
+      "policy/loss": 0.4305807588986923,
+      "step": 946
+    },
+    {
+      "epoch": 0.5050666666666667,
+      "grad_norm": 0.24835757911205292,
+      "learning_rate": 3e-06,
+      "loss": -0.0137,
+      "policy/loss": 0.0601783513729508,
+      "step": 947
+    },
+    {
+      "epoch": 0.5056,
+      "grad_norm": 0.20911452174186707,
+      "learning_rate": 3e-06,
+      "loss": 0.1174,
+      "policy/loss": 0.42783778293030217,
+      "step": 948
+    },
+    {
+      "completion_length": 242.2916717529297,
+      "epoch": 0.5061333333333333,
+      "grad_norm": 0.23054850101470947,
+      "learning_rate": 3e-06,
+      "loss": -0.0724,
+      "policy/loss": 6.930075091560184e-08,
+      "reward": 1.1510416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.4375,
+      "step": 949
+    },
+    {
+      "epoch": 0.5066666666666667,
+      "grad_norm": 0.42095860838890076,
+      "learning_rate": 3e-06,
+      "loss": -0.0771,
+      "policy/loss": 7.168886195785262e-08,
+      "step": 950
+    },
+    {
+      "epoch": 0.5072,
+      "grad_norm": 0.25750046968460083,
+      "learning_rate": 3e-06,
+      "loss": -0.0735,
+      "policy/loss": 6.934275376124788e-08,
+      "step": 951
+    },
+    {
+      "epoch": 0.5077333333333334,
+      "grad_norm": 0.25175127387046814,
+      "learning_rate": 3e-06,
+      "loss": -0.0786,
+      "policy/loss": 7.185191730485485e-08,
+      "step": 952
+    },
+    {
+      "completion_length": 240.5416717529297,
+      "epoch": 0.5082666666666666,
+      "grad_norm": 0.225553497672081,
+      "learning_rate": 3e-06,
+      "loss": 0.0456,
+      "policy/loss": -0.05807493877217951,
+      "reward": 1.3645833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6458333283662796,
+      "step": 953
+    },
+    {
+      "epoch": 0.5088,
+      "grad_norm": 0.22553686797618866,
+      "learning_rate": 3e-06,
+      "loss": -0.0581,
+      "policy/loss": -0.173680173557635,
+      "step": 954
+    },
+    {
+      "epoch": 0.5093333333333333,
+      "grad_norm": 0.21041738986968994,
+      "learning_rate": 3e-06,
+      "loss": 0.0444,
+      "policy/loss": -0.05884294090554221,
+      "step": 955
+    },
+    {
+      "epoch": 0.5098666666666667,
+      "grad_norm": 0.24546588957309723,
+      "learning_rate": 3e-06,
+      "loss": -0.0589,
+      "policy/loss": -0.17324464234918935,
+      "step": 956
+    },
+    {
+      "completion_length": 245.2916717529297,
+      "epoch": 0.5104,
+      "grad_norm": 0.20983846485614777,
+      "learning_rate": 3e-06,
+      "loss": -0.0428,
+      "policy/loss": -0.2441347111016512,
+      "reward": 1.8437500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 1.1875,
+      "step": 957
+    },
+    {
+      "epoch": 0.5109333333333334,
+      "grad_norm": 0.21505817770957947,
+      "learning_rate": 3e-06,
+      "loss": -0.0093,
+      "policy/loss": 0.06295731663703918,
+      "step": 958
+    },
+    {
+      "epoch": 0.5114666666666666,
+      "grad_norm": 0.2088208794593811,
+      "learning_rate": 3e-06,
+      "loss": -0.0447,
+      "policy/loss": -0.24424048885703087,
+      "step": 959
+    },
+    {
+      "epoch": 0.512,
+      "grad_norm": 0.22010356187820435,
+      "learning_rate": 3e-06,
+      "loss": -0.0109,
+      "policy/loss": 0.05864216387271881,
+      "step": 960
+    },
+    {
+      "completion_length": 245.6979217529297,
+      "epoch": 0.5125333333333333,
+      "grad_norm": 0.24914905428886414,
+      "learning_rate": 3e-06,
+      "loss": 0.1134,
+      "policy/loss": 0.006655598324819323,
+      "reward": 1.1041666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333134651184,
+      "rewards/correctness_reward_func_math": 0.3958333358168602,
+      "step": 961
+    },
+    {
+      "epoch": 0.5130666666666667,
+      "grad_norm": 0.22492118179798126,
+      "learning_rate": 3e-06,
+      "loss": 0.039,
+      "policy/loss": -0.005292051664653741,
+      "step": 962
+    },
+    {
+      "epoch": 0.5136,
+      "grad_norm": 0.25017938017845154,
+      "learning_rate": 3e-06,
+      "loss": 0.1127,
+      "policy/loss": 0.005866028169855042,
+      "step": 963
+    },
+    {
+      "epoch": 0.5141333333333333,
+      "grad_norm": 0.22988039255142212,
+      "learning_rate": 3e-06,
+      "loss": 0.0387,
+      "policy/loss": -0.006237712869097578,
+      "step": 964
+    },
+    {
+      "completion_length": 237.4479217529297,
+      "epoch": 0.5146666666666667,
+      "grad_norm": 0.22886259853839874,
+      "learning_rate": 3e-06,
+      "loss": 0.0011,
+      "policy/loss": -0.06568192690610886,
+      "reward": 1.4479167461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 965
+    },
+    {
+      "epoch": 0.5152,
+      "grad_norm": 0.21777096390724182,
+      "learning_rate": 3e-06,
+      "loss": -0.1375,
+      "policy/loss": -0.08871885389089584,
+      "step": 966
+    },
+    {
+      "epoch": 0.5157333333333334,
+      "grad_norm": 0.234333798289299,
+      "learning_rate": 3e-06,
+      "loss": 0.0023,
+      "policy/loss": -0.0666603296995163,
+      "step": 967
+    },
+    {
+      "epoch": 0.5162666666666667,
+      "grad_norm": 0.22467654943466187,
+      "learning_rate": 3e-06,
+      "loss": -0.1388,
+      "policy/loss": -0.09051380306482315,
+      "step": 968
+    },
+    {
+      "completion_length": 247.5104217529297,
+      "epoch": 0.5168,
+      "grad_norm": 0.3710779547691345,
+      "learning_rate": 3e-06,
+      "loss": 0.0938,
+      "policy/loss": 1.1023648838204281e-07,
+      "reward": 1.4479167461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 969
+    },
+    {
+      "epoch": 0.5173333333333333,
+      "grad_norm": 0.24735833704471588,
+      "learning_rate": 3e-06,
+      "loss": 0.0253,
+      "policy/loss": 1.2213714128961328e-07,
+      "step": 970
+    },
+    {
+      "epoch": 0.5178666666666667,
+      "grad_norm": 0.3226465582847595,
+      "learning_rate": 3e-06,
+      "loss": 0.094,
+      "policy/loss": 1.1039291791803407e-07,
+      "step": 971
+    },
+    {
+      "epoch": 0.5184,
+      "grad_norm": 0.2166670709848404,
+      "learning_rate": 3e-06,
+      "loss": 0.0247,
+      "policy/loss": 1.221896575032133e-07,
+      "step": 972
+    },
+    {
+      "completion_length": 235.77083587646484,
+      "epoch": 0.5189333333333334,
+      "grad_norm": 0.23989924788475037,
+      "learning_rate": 3e-06,
+      "loss": -0.0311,
+      "policy/loss": 9.989280869149297e-08,
+      "reward": 1.3958333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7083333432674408,
+      "step": 973
+    },
+    {
+      "epoch": 0.5194666666666666,
+      "grad_norm": 0.2745133936405182,
+      "learning_rate": 3e-06,
+      "loss": 0.1007,
+      "policy/loss": 1.039314128092883e-07,
+      "step": 974
+    },
+    {
+      "epoch": 0.52,
+      "grad_norm": 0.25253233313560486,
+      "learning_rate": 3e-06,
+      "loss": -0.0317,
+      "policy/loss": 9.998196048854879e-08,
+      "step": 975
+    },
+    {
+      "epoch": 0.5205333333333333,
+      "grad_norm": 0.26373863220214844,
+      "learning_rate": 3e-06,
+      "loss": 0.1008,
+      "policy/loss": 1.0381587500774003e-07,
+      "step": 976
+    },
+    {
+      "completion_length": 247.61459350585938,
+      "epoch": 0.5210666666666667,
+      "grad_norm": 0.19619528949260712,
+      "learning_rate": 3e-06,
+      "loss": -0.0778,
+      "policy/loss": 0.2512276873010606,
+      "reward": 1.4375000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7499999850988388,
+      "step": 977
+    },
+    {
+      "epoch": 0.5216,
+      "grad_norm": 0.1992419809103012,
+      "learning_rate": 3e-06,
+      "loss": -0.0462,
+      "policy/loss": 0.07806115539014513,
+      "step": 978
+    },
+    {
+      "epoch": 0.5221333333333333,
+      "grad_norm": 0.21697449684143066,
+      "learning_rate": 3e-06,
+      "loss": -0.0786,
+      "policy/loss": 0.250280748211118,
+      "step": 979
+    },
+    {
+      "epoch": 0.5226666666666666,
+      "grad_norm": 0.26398682594299316,
+      "learning_rate": 3e-06,
+      "loss": -0.047,
+      "policy/loss": 0.07813725564840723,
+      "step": 980
+    },
+    {
+      "completion_length": 235.08334350585938,
+      "epoch": 0.5232,
+      "grad_norm": 0.24352313578128815,
+      "learning_rate": 3e-06,
+      "loss": -0.0107,
+      "policy/loss": -0.0023425668478012085,
+      "reward": 1.875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 1.1875,
+      "step": 981
+    },
+    {
+      "epoch": 0.5237333333333334,
+      "grad_norm": 0.22529570758342743,
+      "learning_rate": 3e-06,
+      "loss": 0.1194,
+      "policy/loss": 0.07880509458482265,
+      "step": 982
+    },
+    {
+      "epoch": 0.5242666666666667,
+      "grad_norm": 0.19501237571239471,
+      "learning_rate": 3e-06,
+      "loss": -0.011,
+      "policy/loss": -0.0013576745986938477,
+      "step": 983
+    },
+    {
+      "epoch": 0.5248,
+      "grad_norm": 0.2270509898662567,
+      "learning_rate": 3e-06,
+      "loss": 0.1173,
+      "policy/loss": 0.07576347328722477,
+      "step": 984
+    },
+    {
+      "completion_length": 241.6354217529297,
+      "epoch": 0.5253333333333333,
+      "grad_norm": 0.2720073461532593,
+      "learning_rate": 3e-06,
+      "loss": 0.0597,
+      "policy/loss": 0.248857615418137,
+      "reward": 1.7604166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 1.0416666567325592,
+      "step": 985
+    },
+    {
+      "epoch": 0.5258666666666667,
+      "grad_norm": 0.21881301701068878,
+      "learning_rate": 3e-06,
+      "loss": 0.04,
+      "policy/loss": -0.27584585402605555,
+      "step": 986
+    },
+    {
+      "epoch": 0.5264,
+      "grad_norm": 0.2547081708908081,
+      "learning_rate": 3e-06,
+      "loss": 0.0589,
+      "policy/loss": 0.2502204457209132,
+      "step": 987
+    },
+    {
+      "epoch": 0.5269333333333334,
+      "grad_norm": 0.21073578298091888,
+      "learning_rate": 3e-06,
+      "loss": 0.0391,
+      "policy/loss": -0.27791935068339413,
+      "step": 988
+    },
+    {
+      "completion_length": 239.14583587646484,
+      "epoch": 0.5274666666666666,
+      "grad_norm": 0.19903874397277832,
+      "learning_rate": 3e-06,
+      "loss": -0.0466,
+      "policy/loss": 1.4028712769231788e-07,
+      "reward": 1.2291666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.5416666716337204,
+      "step": 989
+    },
+    {
+      "epoch": 0.528,
+      "grad_norm": 0.2569504380226135,
+      "learning_rate": 3e-06,
+      "loss": -0.084,
+      "policy/loss": 1.648383118890706e-07,
+      "step": 990
+    },
+    {
+      "epoch": 0.5285333333333333,
+      "grad_norm": 0.2025875747203827,
+      "learning_rate": 3e-06,
+      "loss": -0.0468,
+      "policy/loss": 1.4034367268322967e-07,
+      "step": 991
+    },
+    {
+      "epoch": 0.5290666666666667,
+      "grad_norm": 0.29295557737350464,
+      "learning_rate": 3e-06,
+      "loss": -0.0863,
+      "policy/loss": 1.6484548126527443e-07,
+      "step": 992
+    },
+    {
+      "completion_length": 246.11459350585938,
+      "epoch": 0.5296,
+      "grad_norm": 0.5292813181877136,
+      "learning_rate": 3e-06,
+      "loss": -0.0037,
+      "policy/loss": -0.1253458559513092,
+      "reward": 1.3854166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6666666641831398,
+      "step": 993
+    },
+    {
+      "epoch": 0.5301333333333333,
+      "grad_norm": 0.2076614946126938,
+      "learning_rate": 3e-06,
+      "loss": -0.0838,
+      "policy/loss": -0.031123802065849304,
+      "step": 994
+    },
+    {
+      "epoch": 0.5306666666666666,
+      "grad_norm": 0.27536797523498535,
+      "learning_rate": 3e-06,
+      "loss": -0.004,
+      "policy/loss": -0.12720715999603271,
+      "step": 995
+    },
+    {
+      "epoch": 0.5312,
+      "grad_norm": 0.32765910029411316,
+      "learning_rate": 3e-06,
+      "loss": -0.0841,
+      "policy/loss": -0.032110944390296936,
+      "step": 996
+    },
+    {
+      "completion_length": 243.8541717529297,
+      "epoch": 0.5317333333333333,
+      "grad_norm": 0.2940518260002136,
+      "learning_rate": 3e-06,
+      "loss": 0.0461,
+      "policy/loss": 0.2499336302280426,
+      "reward": 1.421875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.7708333134651184,
+      "step": 997
+    },
+    {
+      "epoch": 0.5322666666666667,
+      "grad_norm": 0.3208288848400116,
+      "learning_rate": 3e-06,
+      "loss": 0.0386,
+      "policy/loss": -0.11505214869976044,
+      "step": 998
+    },
+    {
+      "epoch": 0.5328,
+      "grad_norm": 0.3252948522567749,
+      "learning_rate": 3e-06,
+      "loss": 0.0453,
+      "policy/loss": 0.25333255529403687,
+      "step": 999
+    },
+    {
+      "epoch": 0.5333333333333333,
+      "grad_norm": 0.37325558066368103,
+      "learning_rate": 3e-06,
+      "loss": 0.0367,
+      "policy/loss": -0.11622326821088791,
+      "step": 1000
+    },
+    {
+      "completion_length": 234.5104217529297,
+      "epoch": 0.5338666666666667,
+      "grad_norm": 0.28701651096343994,
+      "learning_rate": 3e-06,
+      "loss": 0.0121,
+      "policy/loss": 0.22830586152426235,
+      "reward": 1.5104166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 1001
+    },
+    {
+      "epoch": 0.5344,
+      "grad_norm": 0.27529391646385193,
+      "learning_rate": 3e-06,
+      "loss": 0.1461,
+      "policy/loss": -0.11696434963283409,
+      "step": 1002
+    },
+    {
+      "epoch": 0.5349333333333334,
+      "grad_norm": 0.2443794459104538,
+      "learning_rate": 3e-06,
+      "loss": 0.0101,
+      "policy/loss": 0.23003473898818783,
+      "step": 1003
+    },
+    {
+      "epoch": 0.5354666666666666,
+      "grad_norm": 0.28580939769744873,
+      "learning_rate": 3e-06,
+      "loss": 0.1436,
+      "policy/loss": -0.11973642749746816,
+      "step": 1004
+    },
+    {
+      "completion_length": 241.02084350585938,
+      "epoch": 0.536,
+      "grad_norm": 0.2631276845932007,
+      "learning_rate": 3e-06,
+      "loss": -0.051,
+      "policy/loss": 1.0102265335376615e-07,
+      "reward": 1.3020833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.5833333283662796,
+      "step": 1005
+    },
+    {
+      "epoch": 0.5365333333333333,
+      "grad_norm": 0.31385183334350586,
+      "learning_rate": 3e-06,
+      "loss": -0.0319,
+      "policy/loss": 1.1670412902731186e-07,
+      "step": 1006
+    },
+    {
+      "epoch": 0.5370666666666667,
+      "grad_norm": 0.24480947852134705,
+      "learning_rate": 3e-06,
+      "loss": -0.0522,
+      "policy/loss": 1.0110415615827151e-07,
+      "step": 1007
+    },
+    {
+      "epoch": 0.5376,
+      "grad_norm": 0.6358599066734314,
+      "learning_rate": 3e-06,
+      "loss": -0.0325,
+      "policy/loss": 1.1685161638297359e-07,
+      "step": 1008
+    },
+    {
+      "completion_length": 236.1666717529297,
+      "epoch": 0.5381333333333334,
+      "grad_norm": 0.24889059364795685,
+      "learning_rate": 3e-06,
+      "loss": 0.0039,
+      "policy/loss": 9.928957922511472e-08,
+      "reward": 1.5989583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.9166666567325592,
+      "step": 1009
+    },
+    {
+      "epoch": 0.5386666666666666,
+      "grad_norm": 0.2386978566646576,
+      "learning_rate": 3e-06,
+      "loss": -0.1339,
+      "policy/loss": 1.129656084231101e-07,
+      "step": 1010
+    },
+    {
+      "epoch": 0.5392,
+      "grad_norm": 0.2525117099285126,
+      "learning_rate": 3e-06,
+      "loss": 0.0031,
+      "policy/loss": 9.924001531658178e-08,
+      "step": 1011
+    },
+    {
+      "epoch": 0.5397333333333333,
+      "grad_norm": 0.23515182733535767,
+      "learning_rate": 3e-06,
+      "loss": -0.1352,
+      "policy/loss": 1.1302743985197594e-07,
+      "step": 1012
+    },
+    {
+      "completion_length": 238.3125,
+      "epoch": 0.5402666666666667,
+      "grad_norm": 0.23697853088378906,
+      "learning_rate": 3e-06,
+      "loss": -0.0498,
+      "policy/loss": 8.162720632753917e-08,
+      "reward": 1.2083333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.4583333283662796,
+      "step": 1013
+    },
+    {
+      "epoch": 0.5408,
+      "grad_norm": 0.19228550791740417,
+      "learning_rate": 3e-06,
+      "loss": -0.1185,
+      "policy/loss": 8.850130939208611e-08,
+      "step": 1014
+    },
+    {
+      "epoch": 0.5413333333333333,
+      "grad_norm": 0.2318982481956482,
+      "learning_rate": 3e-06,
+      "loss": -0.05,
+      "policy/loss": 8.171636878273603e-08,
+      "step": 1015
+    },
+    {
+      "epoch": 0.5418666666666667,
+      "grad_norm": 0.19420520961284637,
+      "learning_rate": 3e-06,
+      "loss": -0.119,
+      "policy/loss": 8.86697684165938e-08,
+      "step": 1016
+    },
+    {
+      "completion_length": 240.55208587646484,
+      "epoch": 0.5424,
+      "grad_norm": 0.16628849506378174,
+      "learning_rate": 3e-06,
+      "loss": 0.0501,
+      "policy/loss": 1.1998836590976225e-07,
+      "reward": 1.3489583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.6666666567325592,
+      "step": 1017
+    },
+    {
+      "epoch": 0.5429333333333334,
+      "grad_norm": 0.17592470347881317,
+      "learning_rate": 3e-06,
+      "loss": -0.0609,
+      "policy/loss": 1.0018320573124129e-07,
+      "step": 1018
+    },
+    {
+      "epoch": 0.5434666666666667,
+      "grad_norm": 0.16194039583206177,
+      "learning_rate": 3e-06,
+      "loss": 0.0497,
+      "policy/loss": 1.2007468086494555e-07,
+      "step": 1019
+    },
+    {
+      "epoch": 0.544,
+      "grad_norm": 0.18532168865203857,
+      "learning_rate": 3e-06,
+      "loss": -0.0619,
+      "policy/loss": 1.0011904549855899e-07,
+      "step": 1020
+    },
+    {
+      "completion_length": 236.8125,
+      "epoch": 0.5445333333333333,
+      "grad_norm": 0.2104501724243164,
+      "learning_rate": 3e-06,
+      "loss": 0.0411,
+      "policy/loss": -0.16797596880543608,
+      "reward": 1.59375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.875,
+      "step": 1021
+    },
+    {
+      "epoch": 0.5450666666666667,
+      "grad_norm": 0.3140813708305359,
+      "learning_rate": 3e-06,
+      "loss": 0.0997,
+      "policy/loss": 0.25070190680461124,
+      "step": 1022
+    },
+    {
+      "epoch": 0.5456,
+      "grad_norm": 0.1966320425271988,
+      "learning_rate": 3e-06,
+      "loss": 0.0399,
+      "policy/loss": -0.1696053808675977,
+      "step": 1023
+    },
+    {
+      "epoch": 0.5461333333333334,
+      "grad_norm": 0.2990897595882416,
+      "learning_rate": 3e-06,
+      "loss": 0.0994,
+      "policy/loss": 0.25138911849875,
+      "step": 1024
+    },
+    {
+      "completion_length": 242.1354217529297,
+      "epoch": 0.5466666666666666,
+      "grad_norm": 0.21298973262310028,
+      "learning_rate": 3e-06,
+      "loss": -0.0647,
+      "policy/loss": -0.2336680037714931,
+      "reward": 1.1354166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.4791666641831398,
+      "step": 1025
+    },
+    {
+      "epoch": 0.5472,
+      "grad_norm": 0.19707328081130981,
+      "learning_rate": 3e-06,
+      "loss": 0.1021,
+      "policy/loss": 0.2710047345366835,
+      "step": 1026
+    },
+    {
+      "epoch": 0.5477333333333333,
+      "grad_norm": 0.2077799141407013,
+      "learning_rate": 3e-06,
+      "loss": -0.0651,
+      "policy/loss": -0.23267579510622305,
+      "step": 1027
+    },
+    {
+      "epoch": 0.5482666666666667,
+      "grad_norm": 0.19753475487232208,
+      "learning_rate": 3e-06,
+      "loss": 0.1006,
+      "policy/loss": 0.2702707927557597,
+      "step": 1028
+    },
+    {
+      "completion_length": 226.28125762939453,
+      "epoch": 0.5488,
+      "grad_norm": 0.2747688591480255,
+      "learning_rate": 3e-06,
+      "loss": 0.0003,
+      "policy/loss": -0.18945443975209564,
+      "reward": 1.5364583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.7916666865348816,
+      "step": 1029
+    },
+    {
+      "epoch": 0.5493333333333333,
+      "grad_norm": 0.27592241764068604,
+      "learning_rate": 3e-06,
+      "loss": -0.0472,
+      "policy/loss": 0.0131684298655248,
+      "step": 1030
+    },
+    {
+      "epoch": 0.5498666666666666,
+      "grad_norm": 0.24127475917339325,
+      "learning_rate": 3e-06,
+      "loss": -0.0011,
+      "policy/loss": -0.1908974533650003,
+      "step": 1031
+    },
+    {
+      "epoch": 0.5504,
+      "grad_norm": 0.2601408362388611,
+      "learning_rate": 3e-06,
+      "loss": -0.0462,
+      "policy/loss": 0.015203020347458462,
+      "step": 1032
+    },
+    {
+      "completion_length": 234.34375,
+      "epoch": 0.5509333333333334,
+      "grad_norm": 1.4884941577911377,
+      "learning_rate": 3e-06,
+      "loss": -0.0751,
+      "policy/loss": -0.10643092443084612,
+      "reward": 1.2552083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.6041666716337204,
+      "step": 1033
+    },
+    {
+      "epoch": 0.5514666666666667,
+      "grad_norm": 0.2342720776796341,
+      "learning_rate": 3e-06,
+      "loss": -0.0173,
+      "policy/loss": -0.2267948141077003,
+      "step": 1034
+    },
+    {
+      "epoch": 0.552,
+      "grad_norm": 0.2036047875881195,
+      "learning_rate": 3e-06,
+      "loss": -0.0756,
+      "policy/loss": -0.10543181646814226,
+      "step": 1035
+    },
+    {
+      "epoch": 0.5525333333333333,
+      "grad_norm": 0.22528256475925446,
+      "learning_rate": 3e-06,
+      "loss": -0.0182,
+      "policy/loss": -0.22676796219459128,
+      "step": 1036
+    },
+    {
+      "completion_length": 238.8229217529297,
+      "epoch": 0.5530666666666667,
+      "grad_norm": 0.2785310745239258,
+      "learning_rate": 3e-06,
+      "loss": -0.0111,
+      "policy/loss": -0.46909667551517487,
+      "reward": 2.036458373069763,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 1.2916666865348816,
+      "step": 1037
+    },
+    {
+      "epoch": 0.5536,
+      "grad_norm": 0.3016909658908844,
+      "learning_rate": 3e-06,
+      "loss": -0.0096,
+      "policy/loss": 0.30071423947811127,
+      "step": 1038
+    },
+    {
+      "epoch": 0.5541333333333334,
+      "grad_norm": 0.2776082754135132,
+      "learning_rate": 3e-06,
+      "loss": -0.0114,
+      "policy/loss": -0.4672619551420212,
+      "step": 1039
+    },
+    {
+      "epoch": 0.5546666666666666,
+      "grad_norm": 0.3841075301170349,
+      "learning_rate": 3e-06,
+      "loss": -0.0116,
+      "policy/loss": 0.3022681623697281,
+      "step": 1040
+    },
+    {
+      "completion_length": 237.87500762939453,
+      "epoch": 0.5552,
+      "grad_norm": 0.2532280683517456,
+      "learning_rate": 3e-06,
+      "loss": 0.1215,
+      "policy/loss": 0.1383376283623008,
+      "reward": 1.6145833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8958333432674408,
+      "step": 1041
+    },
+    {
+      "epoch": 0.5557333333333333,
+      "grad_norm": 0.3061031997203827,
+      "learning_rate": 3e-06,
+      "loss": -0.1025,
+      "policy/loss": 0.013322334264406877,
+      "step": 1042
+    },
+    {
+      "epoch": 0.5562666666666667,
+      "grad_norm": 0.2490016371011734,
+      "learning_rate": 3e-06,
+      "loss": 0.1206,
+      "policy/loss": 0.13753841949998602,
+      "step": 1043
+    },
+    {
+      "epoch": 0.5568,
+      "grad_norm": 0.25609374046325684,
+      "learning_rate": 3e-06,
+      "loss": -0.1032,
+      "policy/loss": 0.014330278661409679,
+      "step": 1044
+    },
+    {
+      "completion_length": 239.05208587646484,
+      "epoch": 0.5573333333333333,
+      "grad_norm": 0.23267598450183868,
+      "learning_rate": 3e-06,
+      "loss": 0.1405,
+      "policy/loss": 1.633521904409463e-07,
+      "reward": 1.5156250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.8333333134651184,
+      "step": 1045
+    },
+    {
+      "epoch": 0.5578666666666666,
+      "grad_norm": 0.2876987159252167,
+      "learning_rate": 3e-06,
+      "loss": -0.0359,
+      "policy/loss": 2.1573213615511122e-07,
+      "step": 1046
+    },
+    {
+      "epoch": 0.5584,
+      "grad_norm": 0.24781765043735504,
+      "learning_rate": 3e-06,
+      "loss": 0.139,
+      "policy/loss": 1.6336228014779408e-07,
+      "step": 1047
+    },
+    {
+      "epoch": 0.5589333333333333,
+      "grad_norm": 0.26283273100852966,
+      "learning_rate": 3e-06,
+      "loss": -0.0374,
+      "policy/loss": 2.1569515951114226e-07,
+      "step": 1048
+    },
+    {
+      "completion_length": 240.06250762939453,
+      "epoch": 0.5594666666666667,
+      "grad_norm": 0.24345384538173676,
+      "learning_rate": 3e-06,
+      "loss": -0.1653,
+      "policy/loss": -0.08255844283459979,
+      "reward": 0.96875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.59375,
+      "rewards/correctness_reward_func_math": 0.375,
+      "step": 1049
+    },
+    {
+      "epoch": 0.56,
+      "grad_norm": 0.3005201816558838,
+      "learning_rate": 3e-06,
+      "loss": 0.0478,
+      "policy/loss": -0.14220914660833728,
+      "step": 1050
+    },
+    {
+      "epoch": 0.5605333333333333,
+      "grad_norm": 0.2111940234899521,
+      "learning_rate": 3e-06,
+      "loss": -0.1665,
+      "policy/loss": -0.08675190507804942,
+      "step": 1051
+    },
+    {
+      "epoch": 0.5610666666666667,
+      "grad_norm": 0.3326101303100586,
+      "learning_rate": 3e-06,
+      "loss": 0.049,
+      "policy/loss": -0.14227822828372894,
+      "step": 1052
+    },
+    {
+      "completion_length": 235.17708587646484,
+      "epoch": 0.5616,
+      "grad_norm": 0.25171133875846863,
+      "learning_rate": 3e-06,
+      "loss": 0.0344,
+      "policy/loss": 0.08271819603435304,
+      "reward": 1.375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 1053
+    },
+    {
+      "epoch": 0.5621333333333334,
+      "grad_norm": 0.2697506546974182,
+      "learning_rate": 3e-06,
+      "loss": 0.0473,
+      "policy/loss": -0.28706942930620016,
+      "step": 1054
+    },
+    {
+      "epoch": 0.5626666666666666,
+      "grad_norm": 0.2481641322374344,
+      "learning_rate": 3e-06,
+      "loss": 0.0338,
+      "policy/loss": 0.08133195601307364,
+      "step": 1055
+    },
+    {
+      "epoch": 0.5632,
+      "grad_norm": 0.29496437311172485,
+      "learning_rate": 3e-06,
+      "loss": 0.0461,
+      "policy/loss": -0.2893417073064981,
+      "step": 1056
+    },
+    {
+      "completion_length": 231.8541717529297,
+      "epoch": 0.5637333333333333,
+      "grad_norm": 0.19672144949436188,
+      "learning_rate": 3e-06,
+      "loss": -0.1116,
+      "policy/loss": -0.2202462950863726,
+      "reward": 1.3854166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 1057
+    },
+    {
+      "epoch": 0.5642666666666667,
+      "grad_norm": 0.3306174874305725,
+      "learning_rate": 3e-06,
+      "loss": -0.238,
+      "policy/loss": -0.03998587863142333,
+      "step": 1058
+    },
+    {
+      "epoch": 0.5648,
+      "grad_norm": 0.2020600140094757,
+      "learning_rate": 3e-06,
+      "loss": -0.1114,
+      "policy/loss": -0.2217502991102691,
+      "step": 1059
+    },
+    {
+      "epoch": 0.5653333333333334,
+      "grad_norm": 0.3105413615703583,
+      "learning_rate": 3e-06,
+      "loss": -0.2385,
+      "policy/loss": -0.04120633228585646,
+      "step": 1060
+    },
+    {
+      "completion_length": 240.8229217529297,
+      "epoch": 0.5658666666666666,
+      "grad_norm": 0.1891690045595169,
+      "learning_rate": 3e-06,
+      "loss": -0.039,
+      "policy/loss": 0.5212567166575184,
+      "reward": 1.4739583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 1061
+    },
+    {
+      "epoch": 0.5664,
+      "grad_norm": 0.17251847684383392,
+      "learning_rate": 3e-06,
+      "loss": -0.089,
+      "policy/loss": -0.043391736205173004,
+      "step": 1062
+    },
+    {
+      "epoch": 0.5669333333333333,
+      "grad_norm": 0.2076432704925537,
+      "learning_rate": 3e-06,
+      "loss": -0.0397,
+      "policy/loss": 0.5207783893692799,
+      "step": 1063
+    },
+    {
+      "epoch": 0.5674666666666667,
+      "grad_norm": 0.1659197211265564,
+      "learning_rate": 3e-06,
+      "loss": -0.0898,
+      "policy/loss": -0.044384585652048614,
+      "step": 1064
+    },
+    {
+      "completion_length": 240.42708587646484,
+      "epoch": 0.568,
+      "grad_norm": 0.25911980867385864,
+      "learning_rate": 3e-06,
+      "loss": -0.1436,
+      "policy/loss": 8.788754612965022e-08,
+      "reward": 1.4375000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 1065
+    },
+    {
+      "epoch": 0.5685333333333333,
+      "grad_norm": 0.4109334349632263,
+      "learning_rate": 3e-06,
+      "loss": 0.0568,
+      "policy/loss": 1.052117610811365e-07,
+      "step": 1066
+    },
+    {
+      "epoch": 0.5690666666666667,
+      "grad_norm": 0.2516782581806183,
+      "learning_rate": 3e-06,
+      "loss": -0.1447,
+      "policy/loss": 8.819944241622579e-08,
+      "step": 1067
+    },
+    {
+      "epoch": 0.5696,
+      "grad_norm": 0.4092002213001251,
+      "learning_rate": 3e-06,
+      "loss": 0.0541,
+      "policy/loss": 1.0538265726722784e-07,
+      "step": 1068
+    },
+    {
+      "completion_length": 240.39583587646484,
+      "epoch": 0.5701333333333334,
+      "grad_norm": 0.3086097836494446,
+      "learning_rate": 3e-06,
+      "loss": -0.2098,
+      "policy/loss": 7.432296555975881e-08,
+      "reward": 1.3697916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.703125,
+      "rewards/correctness_reward_func_math": 0.6666666567325592,
+      "step": 1069
+    },
+    {
+      "epoch": 0.5706666666666667,
+      "grad_norm": 0.2670655846595764,
+      "learning_rate": 3e-06,
+      "loss": -0.0823,
+      "policy/loss": 6.929567319957641e-08,
+      "step": 1070
+    },
+    {
+      "epoch": 0.5712,
+      "grad_norm": 0.31387653946876526,
+      "learning_rate": 3e-06,
+      "loss": -0.2107,
+      "policy/loss": 7.433309612281391e-08,
+      "step": 1071
+    },
+    {
+      "epoch": 0.5717333333333333,
+      "grad_norm": 0.26497143507003784,
+      "learning_rate": 3e-06,
+      "loss": -0.0836,
+      "policy/loss": 6.926772933013581e-08,
+      "step": 1072
+    },
+    {
+      "completion_length": 239.98958587646484,
+      "epoch": 0.5722666666666667,
+      "grad_norm": 0.2643846869468689,
+      "learning_rate": 3e-06,
+      "loss": -0.1092,
+      "policy/loss": -0.312144935131073,
+      "reward": 1.0260416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6302083432674408,
+      "rewards/correctness_reward_func_math": 0.395833320915699,
+      "step": 1073
+    },
+    {
+      "epoch": 0.5728,
+      "grad_norm": 0.2886006832122803,
+      "learning_rate": 3e-06,
+      "loss": 0.0934,
+      "policy/loss": 0.48814964294433594,
+      "step": 1074
+    },
+    {
+      "epoch": 0.5733333333333334,
+      "grad_norm": 0.2534918487071991,
+      "learning_rate": 3e-06,
+      "loss": -0.1114,
+      "policy/loss": -0.3143167495727539,
+      "step": 1075
+    },
+    {
+      "epoch": 0.5738666666666666,
+      "grad_norm": 0.297865629196167,
+      "learning_rate": 3e-06,
+      "loss": 0.0913,
+      "policy/loss": 0.4869459718465805,
+      "step": 1076
+    },
+    {
+      "completion_length": 244.6979217529297,
+      "epoch": 0.5744,
+      "grad_norm": 0.2540026307106018,
+      "learning_rate": 3e-06,
+      "loss": -0.1252,
+      "policy/loss": -0.20804008096456528,
+      "reward": 1.4322916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 1077
+    },
+    {
+      "epoch": 0.5749333333333333,
+      "grad_norm": 0.19379015266895294,
+      "learning_rate": 3e-06,
+      "loss": 0.0037,
+      "policy/loss": -0.1281775743700564,
+      "step": 1078
+    },
+    {
+      "epoch": 0.5754666666666667,
+      "grad_norm": 0.2796920835971832,
+      "learning_rate": 3e-06,
+      "loss": -0.1262,
+      "policy/loss": -0.21304794400930405,
+      "step": 1079
+    },
+    {
+      "epoch": 0.576,
+      "grad_norm": 0.19566576182842255,
+      "learning_rate": 3e-06,
+      "loss": 0.0034,
+      "policy/loss": -0.1266484367661178,
+      "step": 1080
+    },
+    {
+      "completion_length": 242.42708587646484,
+      "epoch": 0.5765333333333333,
+      "grad_norm": 0.19261138141155243,
+      "learning_rate": 3e-06,
+      "loss": 0.0969,
+      "policy/loss": 8.237702253666157e-08,
+      "reward": 1.2187500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.5625,
+      "step": 1081
+    },
+    {
+      "epoch": 0.5770666666666666,
+      "grad_norm": 0.21334519982337952,
+      "learning_rate": 3e-06,
+      "loss": 0.0075,
+      "policy/loss": 8.880402369015883e-08,
+      "step": 1082
+    },
+    {
+      "epoch": 0.5776,
+      "grad_norm": 0.18929511308670044,
+      "learning_rate": 3e-06,
+      "loss": 0.096,
+      "policy/loss": 8.240521687241653e-08,
+      "step": 1083
+    },
+    {
+      "epoch": 0.5781333333333334,
+      "grad_norm": 0.20475240051746368,
+      "learning_rate": 3e-06,
+      "loss": 0.0066,
+      "policy/loss": 8.868038037235237e-08,
+      "step": 1084
+    },
+    {
+      "completion_length": 236.59375762939453,
+      "epoch": 0.5786666666666667,
+      "grad_norm": 0.23762854933738708,
+      "learning_rate": 3e-06,
+      "loss": 0.1113,
+      "policy/loss": 9.331163397519049e-08,
+      "reward": 1.5156250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.7708333432674408,
+      "step": 1085
+    },
+    {
+      "epoch": 0.5792,
+      "grad_norm": 0.20619745552539825,
+      "learning_rate": 3e-06,
+      "loss": -0.0739,
+      "policy/loss": 7.827620507327993e-08,
+      "step": 1086
+    },
+    {
+      "epoch": 0.5797333333333333,
+      "grad_norm": 0.24801963567733765,
+      "learning_rate": 3e-06,
+      "loss": 0.1101,
+      "policy/loss": 9.33686941095857e-08,
+      "step": 1087
+    },
+    {
+      "epoch": 0.5802666666666667,
+      "grad_norm": 0.22085252404212952,
+      "learning_rate": 3e-06,
+      "loss": -0.0751,
+      "policy/loss": 7.823985015420476e-08,
+      "step": 1088
+    },
+    {
+      "completion_length": 243.84375762939453,
+      "epoch": 0.5808,
+      "grad_norm": 0.23641730844974518,
+      "learning_rate": 3e-06,
+      "loss": -0.0933,
+      "policy/loss": 0.0016169429365078258,
+      "reward": 1.7656250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 1.0208333432674408,
+      "step": 1089
+    },
+    {
+      "epoch": 0.5813333333333334,
+      "grad_norm": 0.2043461799621582,
+      "learning_rate": 3e-06,
+      "loss": -0.0183,
+      "policy/loss": -0.03349844817285863,
+      "step": 1090
+    },
+    {
+      "epoch": 0.5818666666666666,
+      "grad_norm": 0.21188397705554962,
+      "learning_rate": 3e-06,
+      "loss": -0.0942,
+      "policy/loss": 0.0019247570299860683,
+      "step": 1091
+    },
+    {
+      "epoch": 0.5824,
+      "grad_norm": 0.20389460027217865,
+      "learning_rate": 3e-06,
+      "loss": -0.0181,
+      "policy/loss": -0.033152692815031415,
+      "step": 1092
+    },
+    {
+      "completion_length": 231.43750762939453,
+      "epoch": 0.5829333333333333,
+      "grad_norm": 0.21489135921001434,
+      "learning_rate": 3e-06,
+      "loss": 0.1388,
+      "policy/loss": 1.3433435697152163e-07,
+      "reward": 1.7187500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 1.0000000298023224,
+      "step": 1093
+    },
+    {
+      "epoch": 0.5834666666666667,
+      "grad_norm": 0.33789703249931335,
+      "learning_rate": 3e-06,
+      "loss": 0.2293,
+      "policy/loss": 1.293653397738126e-07,
+      "step": 1094
+    },
+    {
+      "epoch": 0.584,
+      "grad_norm": 0.2090294510126114,
+      "learning_rate": 3e-06,
+      "loss": 0.1374,
+      "policy/loss": 1.3454791769618168e-07,
+      "step": 1095
+    },
+    {
+      "epoch": 0.5845333333333333,
+      "grad_norm": 0.40329933166503906,
+      "learning_rate": 3e-06,
+      "loss": 0.2254,
+      "policy/loss": 1.295381863997136e-07,
+      "step": 1096
+    },
+    {
+      "completion_length": 240.0104217529297,
+      "epoch": 0.5850666666666666,
+      "grad_norm": 0.2621570825576782,
+      "learning_rate": 3e-06,
+      "loss": -0.0018,
+      "policy/loss": 0.29689645767211914,
+      "reward": 1.5833333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.8333333134651184,
+      "step": 1097
+    },
+    {
+      "epoch": 0.5856,
+      "grad_norm": 0.2344721108675003,
+      "learning_rate": 3e-06,
+      "loss": -0.1855,
+      "policy/loss": -0.5087320506572723,
+      "step": 1098
+    },
+    {
+      "epoch": 0.5861333333333333,
+      "grad_norm": 0.27209749817848206,
+      "learning_rate": 3e-06,
+      "loss": -0.0028,
+      "policy/loss": 0.29649877175688744,
+      "step": 1099
+    },
+    {
+      "epoch": 0.5866666666666667,
+      "grad_norm": 0.2384234368801117,
+      "learning_rate": 3e-06,
+      "loss": -0.1861,
+      "policy/loss": -0.5102093368768692,
+      "step": 1100
+    },
+    {
+      "completion_length": 237.28125762939453,
+      "epoch": 0.5872,
+      "grad_norm": 0.24861721694469452,
+      "learning_rate": 3e-06,
+      "loss": -0.0182,
+      "policy/loss": 1.0932943084185354e-07,
+      "reward": 1.625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.9375,
+      "step": 1101
+    },
+    {
+      "epoch": 0.5877333333333333,
+      "grad_norm": 0.26820749044418335,
+      "learning_rate": 3e-06,
+      "loss": -0.0616,
+      "policy/loss": 1.2616147770927455e-07,
+      "step": 1102
+    },
+    {
+      "epoch": 0.5882666666666667,
+      "grad_norm": 0.25454849004745483,
+      "learning_rate": 3e-06,
+      "loss": -0.0182,
+      "policy/loss": 1.093319497158518e-07,
+      "step": 1103
+    },
+    {
+      "epoch": 0.5888,
+      "grad_norm": 0.22327356040477753,
+      "learning_rate": 3e-06,
+      "loss": -0.0617,
+      "policy/loss": 1.2611810973339743e-07,
+      "step": 1104
+    },
+    {
+      "completion_length": 242.9375,
+      "epoch": 0.5893333333333334,
+      "grad_norm": 0.14943358302116394,
+      "learning_rate": 3e-06,
+      "loss": 0.1698,
+      "policy/loss": 0.4679849445819855,
+      "reward": 1.0989583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.4166666567325592,
+      "step": 1105
+    },
+    {
+      "epoch": 0.5898666666666667,
+      "grad_norm": 0.1699458509683609,
+      "learning_rate": 3e-06,
+      "loss": 0.074,
+      "policy/loss": -0.11679781228303909,
+      "step": 1106
+    },
+    {
+      "epoch": 0.5904,
+      "grad_norm": 0.15052767097949982,
+      "learning_rate": 3e-06,
+      "loss": 0.1698,
+      "policy/loss": 0.46735602617263794,
+      "step": 1107
+    },
+    {
+      "epoch": 0.5909333333333333,
+      "grad_norm": 0.13748140633106232,
+      "learning_rate": 3e-06,
+      "loss": 0.0729,
+      "policy/loss": -0.12038545310497284,
+      "step": 1108
+    },
+    {
+      "completion_length": 234.52083587646484,
+      "epoch": 0.5914666666666667,
+      "grad_norm": 0.22554421424865723,
+      "learning_rate": 3e-06,
+      "loss": -0.0772,
+      "policy/loss": 0.02980300168638017,
+      "reward": 1.375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 1109
+    },
+    {
+      "epoch": 0.592,
+      "grad_norm": 0.3106343746185303,
+      "learning_rate": 3e-06,
+      "loss": 0.077,
+      "policy/loss": -0.09949317109717981,
+      "step": 1110
+    },
+    {
+      "epoch": 0.5925333333333334,
+      "grad_norm": 0.22968870401382446,
+      "learning_rate": 3e-06,
+      "loss": -0.0785,
+      "policy/loss": 0.027467727173133838,
+      "step": 1111
+    },
+    {
+      "epoch": 0.5930666666666666,
+      "grad_norm": 0.26202383637428284,
+      "learning_rate": 3e-06,
+      "loss": 0.0759,
+      "policy/loss": -0.10375128221235919,
+      "step": 1112
+    },
+    {
+      "completion_length": 244.4479217529297,
+      "epoch": 0.5936,
+      "grad_norm": 0.24268388748168945,
+      "learning_rate": 3e-06,
+      "loss": -0.0453,
+      "policy/loss": 8.053249445083566e-08,
+      "reward": 1.1458333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.4583333432674408,
+      "step": 1113
+    },
+    {
+      "epoch": 0.5941333333333333,
+      "grad_norm": 0.25465697050094604,
+      "learning_rate": 3e-06,
+      "loss": 0.0179,
+      "policy/loss": 8.55333013305426e-08,
+      "step": 1114
+    },
+    {
+      "epoch": 0.5946666666666667,
+      "grad_norm": 0.19510845839977264,
+      "learning_rate": 3e-06,
+      "loss": -0.0453,
+      "policy/loss": 8.073552137943807e-08,
+      "step": 1115
+    },
+    {
+      "epoch": 0.5952,
+      "grad_norm": 1.7022886276245117,
+      "learning_rate": 3e-06,
+      "loss": 0.018,
+      "policy/loss": 8.564737896676888e-08,
+      "step": 1116
+    },
+    {
+      "completion_length": 210.5104217529297,
+      "epoch": 0.5957333333333333,
+      "grad_norm": 0.26655682921409607,
+      "learning_rate": 3e-06,
+      "loss": 0.0006,
+      "policy/loss": 0.3348994499905391,
+      "reward": 1.5520833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6770833432674408,
+      "rewards/correctness_reward_func_math": 0.875,
+      "step": 1117
+    },
+    {
+      "epoch": 0.5962666666666666,
+      "grad_norm": 0.28897544741630554,
+      "learning_rate": 3e-06,
+      "loss": -0.1579,
+      "policy/loss": -0.20471074695078428,
+      "step": 1118
+    },
+    {
+      "epoch": 0.5968,
+      "grad_norm": 0.26261106133461,
+      "learning_rate": 3e-06,
+      "loss": -0.0021,
+      "policy/loss": 0.3329025453509473,
+      "step": 1119
+    },
+    {
+      "epoch": 0.5973333333333334,
+      "grad_norm": 0.3160240650177002,
+      "learning_rate": 3e-06,
+      "loss": -0.1583,
+      "policy/loss": -0.20595332488591467,
+      "step": 1120
+    },
+    {
+      "completion_length": 251.2291717529297,
+      "epoch": 0.5978666666666667,
+      "grad_norm": 0.26995232701301575,
+      "learning_rate": 3e-06,
+      "loss": 0.1186,
+      "policy/loss": 0.4150647727850174,
+      "reward": 1.4583333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.7083333134651184,
+      "step": 1121
+    },
+    {
+      "epoch": 0.5984,
+      "grad_norm": 0.23646031320095062,
+      "learning_rate": 3e-06,
+      "loss": -0.0327,
+      "policy/loss": 0.09226110531781728,
+      "step": 1122
+    },
+    {
+      "epoch": 0.5989333333333333,
+      "grad_norm": 0.32439541816711426,
+      "learning_rate": 3e-06,
+      "loss": 0.1172,
+      "policy/loss": 0.4139539209724177,
+      "step": 1123
+    },
+    {
+      "epoch": 0.5994666666666667,
+      "grad_norm": 0.25209590792655945,
+      "learning_rate": 3e-06,
+      "loss": -0.0335,
+      "policy/loss": 0.09095879605464141,
+      "step": 1124
+    },
+    {
+      "completion_length": 242.50000762939453,
+      "epoch": 0.6,
+      "grad_norm": 0.20991891622543335,
+      "learning_rate": 3e-06,
+      "loss": -0.0054,
+      "policy/loss": 0.04733693676923423,
+      "reward": 1.34375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 1125
+    },
+    {
+      "epoch": 0.6005333333333334,
+      "grad_norm": 0.21295857429504395,
+      "learning_rate": 3e-06,
+      "loss": 0.0255,
+      "policy/loss": -0.04654382022320647,
+      "step": 1126
+    },
+    {
+      "epoch": 0.6010666666666666,
+      "grad_norm": 0.18752379715442657,
+      "learning_rate": 3e-06,
+      "loss": -0.0058,
+      "policy/loss": 0.049209748138348886,
+      "step": 1127
+    },
+    {
+      "epoch": 0.6016,
+      "grad_norm": 0.20894844830036163,
+      "learning_rate": 3e-06,
+      "loss": 0.0243,
+      "policy/loss": -0.0451328589484703,
+      "step": 1128
+    },
+    {
+      "completion_length": 249.5104217529297,
+      "epoch": 0.6021333333333333,
+      "grad_norm": 0.2442057579755783,
+      "learning_rate": 3e-06,
+      "loss": 0.0327,
+      "policy/loss": 6.628616233683715e-08,
+      "reward": 1.6458333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.8958333432674408,
+      "step": 1129
+    },
+    {
+      "epoch": 0.6026666666666667,
+      "grad_norm": 0.2648386061191559,
+      "learning_rate": 3e-06,
+      "loss": 0.1622,
+      "policy/loss": 8.255592121031441e-08,
+      "step": 1130
+    },
+    {
+      "epoch": 0.6032,
+      "grad_norm": 0.24203091859817505,
+      "learning_rate": 3e-06,
+      "loss": 0.0321,
+      "policy/loss": 6.624285209255731e-08,
+      "step": 1131
+    },
+    {
+      "epoch": 0.6037333333333333,
+      "grad_norm": 0.2657654583454132,
+      "learning_rate": 3e-06,
+      "loss": 0.1614,
+      "policy/loss": 8.262020134708337e-08,
+      "step": 1132
+    },
+    {
+      "completion_length": 237.4166717529297,
+      "epoch": 0.6042666666666666,
+      "grad_norm": 0.21988150477409363,
+      "learning_rate": 3e-06,
+      "loss": 0.007,
+      "policy/loss": 0.203029066324234,
+      "reward": 1.3229166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6041666716337204,
+      "step": 1133
+    },
+    {
+      "epoch": 0.6048,
+      "grad_norm": 0.2585243582725525,
+      "learning_rate": 3e-06,
+      "loss": 0.1474,
+      "policy/loss": 0.7516279518604279,
+      "step": 1134
+    },
+    {
+      "epoch": 0.6053333333333333,
+      "grad_norm": 0.22780799865722656,
+      "learning_rate": 3e-06,
+      "loss": 0.0059,
+      "policy/loss": 0.19673515111207962,
+      "step": 1135
+    },
+    {
+      "epoch": 0.6058666666666667,
+      "grad_norm": 0.2521998882293701,
+      "learning_rate": 3e-06,
+      "loss": 0.1463,
+      "policy/loss": 0.7474498450756073,
+      "step": 1136
+    },
+    {
+      "completion_length": 240.73958587646484,
+      "epoch": 0.6064,
+      "grad_norm": 0.2581980228424072,
+      "learning_rate": 3e-06,
+      "loss": 0.0301,
+      "policy/loss": -0.19571781864565452,
+      "reward": 1.8333333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 1.0833333134651184,
+      "step": 1137
+    },
+    {
+      "epoch": 0.6069333333333333,
+      "grad_norm": 0.23046258091926575,
+      "learning_rate": 3e-06,
+      "loss": -0.0836,
+      "policy/loss": 0.04126867239541809,
+      "step": 1138
+    },
+    {
+      "epoch": 0.6074666666666667,
+      "grad_norm": 0.21486082673072815,
+      "learning_rate": 3e-06,
+      "loss": 0.0286,
+      "policy/loss": -0.1967964243305289,
+      "step": 1139
+    },
+    {
+      "epoch": 0.608,
+      "grad_norm": 0.20741426944732666,
+      "learning_rate": 3e-06,
+      "loss": -0.0848,
+      "policy/loss": 0.03988793829110193,
+      "step": 1140
+    },
+    {
+      "completion_length": 239.7916717529297,
+      "epoch": 0.6085333333333334,
+      "grad_norm": 0.26318445801734924,
+      "learning_rate": 3e-06,
+      "loss": -0.0895,
+      "policy/loss": 2.0086613261582897e-07,
+      "reward": 1.5729166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8541666567325592,
+      "step": 1141
+    },
+    {
+      "epoch": 0.6090666666666666,
+      "grad_norm": 0.2630581855773926,
+      "learning_rate": 3e-06,
+      "loss": 0.0814,
+      "policy/loss": 1.817282111460372e-07,
+      "step": 1142
+    },
+    {
+      "epoch": 0.6096,
+      "grad_norm": 0.27818697690963745,
+      "learning_rate": 3e-06,
+      "loss": -0.0896,
+      "policy/loss": 2.0089147767521354e-07,
+      "step": 1143
+    },
+    {
+      "epoch": 0.6101333333333333,
+      "grad_norm": 0.2743080258369446,
+      "learning_rate": 3e-06,
+      "loss": 0.0801,
+      "policy/loss": 1.8173582816416456e-07,
+      "step": 1144
+    },
+    {
+      "completion_length": 236.6979217529297,
+      "epoch": 0.6106666666666667,
+      "grad_norm": 0.21165123581886292,
+      "learning_rate": 3e-06,
+      "loss": 0.1797,
+      "policy/loss": 0.3771901034652707,
+      "reward": 1.3229166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6041666567325592,
+      "step": 1145
+    },
+    {
+      "epoch": 0.6112,
+      "grad_norm": 0.23841586709022522,
+      "learning_rate": 3e-06,
+      "loss": 0.0694,
+      "policy/loss": -0.016432657438496534,
+      "step": 1146
+    },
+    {
+      "epoch": 0.6117333333333334,
+      "grad_norm": 0.19971394538879395,
+      "learning_rate": 3e-06,
+      "loss": 0.1787,
+      "policy/loss": 0.37619086139909186,
+      "step": 1147
+    },
+    {
+      "epoch": 0.6122666666666666,
+      "grad_norm": 0.20380806922912598,
+      "learning_rate": 3e-06,
+      "loss": 0.0687,
+      "policy/loss": -0.014887463858784145,
+      "step": 1148
+    },
+    {
+      "completion_length": 245.86459350585938,
+      "epoch": 0.6128,
+      "grad_norm": 0.1936064213514328,
+      "learning_rate": 3e-06,
+      "loss": 0.0167,
+      "policy/loss": 3.904172274360462e-08,
+      "reward": 1.3333333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 1149
+    },
+    {
+      "epoch": 0.6133333333333333,
+      "grad_norm": 0.2222164422273636,
+      "learning_rate": 3e-06,
+      "loss": 0.0795,
+      "policy/loss": 2.2232589103055034e-08,
+      "step": 1150
+    },
+    {
+      "epoch": 0.6138666666666667,
+      "grad_norm": 0.18869492411613464,
+      "learning_rate": 3e-06,
+      "loss": 0.0158,
+      "policy/loss": 3.903859990828096e-08,
+      "step": 1151
+    },
+    {
+      "epoch": 0.6144,
+      "grad_norm": 0.22030827403068542,
+      "learning_rate": 3e-06,
+      "loss": 0.0788,
+      "policy/loss": 2.2202259586379114e-08,
+      "step": 1152
+    },
+    {
+      "completion_length": 230.12500762939453,
+      "epoch": 0.6149333333333333,
+      "grad_norm": 0.2534099221229553,
+      "learning_rate": 3e-06,
+      "loss": -0.0139,
+      "policy/loss": 0.4958872790704305,
+      "reward": 1.2395833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.5208333283662796,
+      "step": 1153
+    },
+    {
+      "epoch": 0.6154666666666667,
+      "grad_norm": 0.3329208791255951,
+      "learning_rate": 3e-06,
+      "loss": -0.0432,
+      "policy/loss": -0.32933374481268984,
+      "step": 1154
+    },
+    {
+      "epoch": 0.616,
+      "grad_norm": 0.24961692094802856,
+      "learning_rate": 3e-06,
+      "loss": -0.0145,
+      "policy/loss": 0.488990276693972,
+      "step": 1155
+    },
+    {
+      "epoch": 0.6165333333333334,
+      "grad_norm": 0.3947431743144989,
+      "learning_rate": 3e-06,
+      "loss": -0.0445,
+      "policy/loss": -0.3385134561841845,
+      "step": 1156
+    },
+    {
+      "completion_length": 236.3541717529297,
+      "epoch": 0.6170666666666667,
+      "grad_norm": 0.25842124223709106,
+      "learning_rate": 3e-06,
+      "loss": -0.0938,
+      "policy/loss": -0.18519249859161135,
+      "reward": 1.625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.9375,
+      "step": 1157
+    },
+    {
+      "epoch": 0.6176,
+      "grad_norm": 0.37843990325927734,
+      "learning_rate": 3e-06,
+      "loss": -0.0628,
+      "policy/loss": -0.22517997366391995,
+      "step": 1158
+    },
+    {
+      "epoch": 0.6181333333333333,
+      "grad_norm": 0.24169635772705078,
+      "learning_rate": 3e-06,
+      "loss": -0.0935,
+      "policy/loss": -0.18138520726943597,
+      "step": 1159
+    },
+    {
+      "epoch": 0.6186666666666667,
+      "grad_norm": 0.2952207624912262,
+      "learning_rate": 3e-06,
+      "loss": -0.0634,
+      "policy/loss": -0.22272080514786552,
+      "step": 1160
+    },
+    {
+      "completion_length": 230.71875,
+      "epoch": 0.6192,
+      "grad_norm": 0.25701338052749634,
+      "learning_rate": 3e-06,
+      "loss": 0.1414,
+      "policy/loss": 0.4353724792599678,
+      "reward": 1.4895833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.8333333134651184,
+      "step": 1161
+    },
+    {
+      "epoch": 0.6197333333333334,
+      "grad_norm": 0.2555907368659973,
+      "learning_rate": 3e-06,
+      "loss": 0.0486,
+      "policy/loss": -0.2702684998512268,
+      "step": 1162
+    },
+    {
+      "epoch": 0.6202666666666666,
+      "grad_norm": 0.2435457557439804,
+      "learning_rate": 3e-06,
+      "loss": 0.1396,
+      "policy/loss": 0.42847638577222824,
+      "step": 1163
+    },
+    {
+      "epoch": 0.6208,
+      "grad_norm": 0.245192751288414,
+      "learning_rate": 3e-06,
+      "loss": 0.0476,
+      "policy/loss": -0.2731260433793068,
+      "step": 1164
+    },
+    {
+      "completion_length": 227.8854217529297,
+      "epoch": 0.6213333333333333,
+      "grad_norm": 0.20709551870822906,
+      "learning_rate": 3e-06,
+      "loss": 0.0769,
+      "policy/loss": 1.460547061071793e-07,
+      "reward": 1.5000000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.8124999701976776,
+      "step": 1165
+    },
+    {
+      "epoch": 0.6218666666666667,
+      "grad_norm": 0.2534171938896179,
+      "learning_rate": 3e-06,
+      "loss": -0.1584,
+      "policy/loss": 1.1297254420838954e-07,
+      "step": 1166
+    },
+    {
+      "epoch": 0.6224,
+      "grad_norm": 0.20226909220218658,
+      "learning_rate": 3e-06,
+      "loss": 0.0773,
+      "policy/loss": 1.460738179304144e-07,
+      "step": 1167
+    },
+    {
+      "epoch": 0.6229333333333333,
+      "grad_norm": 0.20140491425991058,
+      "learning_rate": 3e-06,
+      "loss": -0.1598,
+      "policy/loss": 1.131186762037828e-07,
+      "step": 1168
+    },
+    {
+      "completion_length": 244.40625762939453,
+      "epoch": 0.6234666666666666,
+      "grad_norm": 0.24029666185379028,
+      "learning_rate": 3e-06,
+      "loss": -0.0682,
+      "policy/loss": 1.0501977953936148e-07,
+      "reward": 1.4583333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.7083333432674408,
+      "step": 1169
+    },
+    {
+      "epoch": 0.624,
+      "grad_norm": 0.27587762475013733,
+      "learning_rate": 3e-06,
+      "loss": -0.0727,
+      "policy/loss": 8.382767902048727e-08,
+      "step": 1170
+    },
+    {
+      "epoch": 0.6245333333333334,
+      "grad_norm": 0.2398974597454071,
+      "learning_rate": 3e-06,
+      "loss": -0.0695,
+      "policy/loss": 1.0502969161052533e-07,
+      "step": 1171
+    },
+    {
+      "epoch": 0.6250666666666667,
+      "grad_norm": 0.27171656489372253,
+      "learning_rate": 3e-06,
+      "loss": -0.0752,
+      "policy/loss": 8.400022721843925e-08,
+      "step": 1172
+    },
+    {
+      "completion_length": 234.0104217529297,
+      "epoch": 0.6256,
+      "grad_norm": 0.2684139907360077,
+      "learning_rate": 3e-06,
+      "loss": 0.1015,
+      "policy/loss": 0.1187972468429308,
+      "reward": 1.4062500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 1173
+    },
+    {
+      "epoch": 0.6261333333333333,
+      "grad_norm": 0.23941899836063385,
+      "learning_rate": 3e-06,
+      "loss": -0.1311,
+      "policy/loss": -0.005173641083068503,
+      "step": 1174
+    },
+    {
+      "epoch": 0.6266666666666667,
+      "grad_norm": 0.285382479429245,
+      "learning_rate": 3e-06,
+      "loss": 0.1006,
+      "policy/loss": 0.11861771005133548,
+      "step": 1175
+    },
+    {
+      "epoch": 0.6272,
+      "grad_norm": 0.2540619671344757,
+      "learning_rate": 3e-06,
+      "loss": -0.1322,
+      "policy/loss": -0.00609805787755846,
+      "step": 1176
+    },
+    {
+      "completion_length": 239.40625762939453,
+      "epoch": 0.6277333333333334,
+      "grad_norm": 0.2902393341064453,
+      "learning_rate": 3e-06,
+      "loss": 0.0537,
+      "policy/loss": 0.036605130390888974,
+      "reward": 1.4531250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.703125,
+      "rewards/correctness_reward_func_math": 0.7499999850988388,
+      "step": 1177
+    },
+    {
+      "epoch": 0.6282666666666666,
+      "grad_norm": 0.23668821156024933,
+      "learning_rate": 3e-06,
+      "loss": 0.065,
+      "policy/loss": -0.009462679856699907,
+      "step": 1178
+    },
+    {
+      "epoch": 0.6288,
+      "grad_norm": 0.25907593965530396,
+      "learning_rate": 3e-06,
+      "loss": 0.0535,
+      "policy/loss": 0.036789808045416095,
+      "step": 1179
+    },
+    {
+      "epoch": 0.6293333333333333,
+      "grad_norm": 0.24608160555362701,
+      "learning_rate": 3e-06,
+      "loss": 0.0633,
+      "policy/loss": -0.009837079962327522,
+      "step": 1180
+    },
+    {
+      "completion_length": 243.37500762939453,
+      "epoch": 0.6298666666666667,
+      "grad_norm": 0.22873665392398834,
+      "learning_rate": 3e-06,
+      "loss": -0.1313,
+      "policy/loss": -0.12444701045751572,
+      "reward": 1.1770833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.4583333283662796,
+      "step": 1181
+    },
+    {
+      "epoch": 0.6304,
+      "grad_norm": 0.3031920790672302,
+      "learning_rate": 3e-06,
+      "loss": 0.0008,
+      "policy/loss": -0.1472889706492424,
+      "step": 1182
+    },
+    {
+      "epoch": 0.6309333333333333,
+      "grad_norm": 0.23333390057086945,
+      "learning_rate": 3e-06,
+      "loss": -0.1312,
+      "policy/loss": -0.1251293271780014,
+      "step": 1183
+    },
+    {
+      "epoch": 0.6314666666666666,
+      "grad_norm": 0.24615247547626495,
+      "learning_rate": 3e-06,
+      "loss": -0.0004,
+      "policy/loss": -0.14634855091571808,
+      "step": 1184
+    },
+    {
+      "completion_length": 235.03125762939453,
+      "epoch": 0.632,
+      "grad_norm": 0.3658653497695923,
+      "learning_rate": 3e-06,
+      "loss": -0.0075,
+      "policy/loss": 0.03981719667683592,
+      "reward": 1.4583333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333432674408,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 1185
+    },
+    {
+      "epoch": 0.6325333333333333,
+      "grad_norm": 0.27482327818870544,
+      "learning_rate": 3e-06,
+      "loss": 0.0754,
+      "policy/loss": -0.4314771907913375,
+      "step": 1186
+    },
+    {
+      "epoch": 0.6330666666666667,
+      "grad_norm": 0.2881772220134735,
+      "learning_rate": 3e-06,
+      "loss": -0.0103,
+      "policy/loss": 0.037839533168408934,
+      "step": 1187
+    },
+    {
+      "epoch": 0.6336,
+      "grad_norm": 0.2864757180213928,
+      "learning_rate": 3e-06,
+      "loss": 0.0739,
+      "policy/loss": -0.43310398027467656,
+      "step": 1188
+    },
+    {
+      "completion_length": 238.77083587646484,
+      "epoch": 0.6341333333333333,
+      "grad_norm": 0.3057038187980652,
+      "learning_rate": 3e-06,
+      "loss": 0.0364,
+      "policy/loss": -0.05363292695393085,
+      "reward": 1.1250000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.3750000074505806,
+      "step": 1189
+    },
+    {
+      "epoch": 0.6346666666666667,
+      "grad_norm": 0.25765910744667053,
+      "learning_rate": 3e-06,
+      "loss": 0.1178,
+      "policy/loss": 0.10136215351651856,
+      "step": 1190
+    },
+    {
+      "epoch": 0.6352,
+      "grad_norm": 0.32676005363464355,
+      "learning_rate": 3e-06,
+      "loss": 0.0381,
+      "policy/loss": -0.05458498444957627,
+      "step": 1191
+    },
+    {
+      "epoch": 0.6357333333333334,
+      "grad_norm": 0.24460244178771973,
+      "learning_rate": 3e-06,
+      "loss": 0.1156,
+      "policy/loss": 0.098913766057251,
+      "step": 1192
+    },
+    {
+      "completion_length": 221.40625,
+      "epoch": 0.6362666666666666,
+      "grad_norm": 0.20666007697582245,
+      "learning_rate": 3e-06,
+      "loss": 0.1191,
+      "policy/loss": -0.039536610135566974,
+      "reward": 1.0625000298023224,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.4374999925494194,
+      "step": 1193
+    },
+    {
+      "epoch": 0.6368,
+      "grad_norm": 0.2228882610797882,
+      "learning_rate": 3e-06,
+      "loss": 0.0885,
+      "policy/loss": -0.7725559828479742,
+      "step": 1194
+    },
+    {
+      "epoch": 0.6373333333333333,
+      "grad_norm": 0.20344020426273346,
+      "learning_rate": 3e-06,
+      "loss": 0.1179,
+      "policy/loss": -0.042424924654699936,
+      "step": 1195
+    },
+    {
+      "epoch": 0.6378666666666667,
+      "grad_norm": 0.2214038223028183,
+      "learning_rate": 3e-06,
+      "loss": 0.0876,
+      "policy/loss": -0.7734817621755532,
+      "step": 1196
+    },
+    {
+      "completion_length": 224.67708587646484,
+      "epoch": 0.6384,
+      "grad_norm": 0.21439670026302338,
+      "learning_rate": 3e-06,
+      "loss": -0.0693,
+      "policy/loss": -0.25569235609025043,
+      "reward": 1.703125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 1.0208333134651184,
+      "step": 1197
+    },
+    {
+      "epoch": 0.6389333333333334,
+      "grad_norm": 0.19264715909957886,
+      "learning_rate": 3e-06,
+      "loss": 0.013,
+      "policy/loss": -0.01649798785367551,
+      "step": 1198
+    },
+    {
+      "epoch": 0.6394666666666666,
+      "grad_norm": 0.22219102084636688,
+      "learning_rate": 3e-06,
+      "loss": -0.0701,
+      "policy/loss": -0.26071035179317903,
+      "step": 1199
+    },
+    {
+      "epoch": 0.64,
+      "grad_norm": 0.19924719631671906,
+      "learning_rate": 3e-06,
+      "loss": 0.0123,
+      "policy/loss": -0.016001367769327146,
+      "step": 1200
+    },
+    {
+      "completion_length": 235.0,
+      "epoch": 0.6405333333333333,
+      "grad_norm": 0.23097269237041473,
+      "learning_rate": 3e-06,
+      "loss": 0.073,
+      "policy/loss": -0.29465848958867014,
+      "reward": 1.1718750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5885416567325592,
+      "rewards/correctness_reward_func_math": 0.5833333283662796,
+      "step": 1201
+    },
+    {
+      "epoch": 0.6410666666666667,
+      "grad_norm": 0.3544195294380188,
+      "learning_rate": 3e-06,
+      "loss": 0.0479,
+      "policy/loss": -0.4733755154871915,
+      "step": 1202
+    },
+    {
+      "epoch": 0.6416,
+      "grad_norm": 0.222794309258461,
+      "learning_rate": 3e-06,
+      "loss": 0.0725,
+      "policy/loss": -0.3009768203933092,
+      "step": 1203
+    },
+    {
+      "epoch": 0.6421333333333333,
+      "grad_norm": 0.3245648443698883,
+      "learning_rate": 3e-06,
+      "loss": 0.047,
+      "policy/loss": -0.4728361530345637,
+      "step": 1204
+    },
+    {
+      "completion_length": 245.4791717529297,
+      "epoch": 0.6426666666666667,
+      "grad_norm": 0.24657803773880005,
+      "learning_rate": 3e-06,
+      "loss": -0.092,
+      "policy/loss": 0.05065229229743551,
+      "reward": 1.2083333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.5208333283662796,
+      "step": 1205
+    },
+    {
+      "epoch": 0.6432,
+      "grad_norm": 0.2904762029647827,
+      "learning_rate": 3e-06,
+      "loss": -0.0238,
+      "policy/loss": -0.2239388095111181,
+      "step": 1206
+    },
+    {
+      "epoch": 0.6437333333333334,
+      "grad_norm": 0.22979006171226501,
+      "learning_rate": 3e-06,
+      "loss": -0.094,
+      "policy/loss": 0.04858656291676766,
+      "step": 1207
+    },
+    {
+      "epoch": 0.6442666666666667,
+      "grad_norm": 0.28846174478530884,
+      "learning_rate": 3e-06,
+      "loss": -0.025,
+      "policy/loss": -0.23241464950861968,
+      "step": 1208
+    },
+    {
+      "completion_length": 224.1666717529297,
+      "epoch": 0.6448,
+      "grad_norm": 0.3817181885242462,
+      "learning_rate": 3e-06,
+      "loss": 0.1281,
+      "policy/loss": 0.3957238532602787,
+      "reward": 1.8906250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 1.1458333432674408,
+      "step": 1209
+    },
+    {
+      "epoch": 0.6453333333333333,
+      "grad_norm": 0.27970924973487854,
+      "learning_rate": 3e-06,
+      "loss": -0.0984,
+      "policy/loss": -0.08764768578112125,
+      "step": 1210
+    },
+    {
+      "epoch": 0.6458666666666667,
+      "grad_norm": 0.34499114751815796,
+      "learning_rate": 3e-06,
+      "loss": 0.1271,
+      "policy/loss": 0.38600405864417553,
+      "step": 1211
+    },
+    {
+      "epoch": 0.6464,
+      "grad_norm": 0.3371144235134125,
+      "learning_rate": 3e-06,
+      "loss": -0.1006,
+      "policy/loss": -0.09405018761754036,
+      "step": 1212
+    },
+    {
+      "completion_length": 230.89583587646484,
+      "epoch": 0.6469333333333334,
+      "grad_norm": 0.23775805532932281,
+      "learning_rate": 3e-06,
+      "loss": -0.0418,
+      "policy/loss": -0.031080961227416992,
+      "reward": 1.3645833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6458333283662796,
+      "step": 1213
+    },
+    {
+      "epoch": 0.6474666666666666,
+      "grad_norm": 0.2666676342487335,
+      "learning_rate": 3e-06,
+      "loss": -0.0356,
+      "policy/loss": 0.03359028697013855,
+      "step": 1214
+    },
+    {
+      "epoch": 0.648,
+      "grad_norm": 0.2766290605068207,
+      "learning_rate": 3e-06,
+      "loss": -0.0435,
+      "policy/loss": -0.03169085085391998,
+      "step": 1215
+    },
+    {
+      "epoch": 0.6485333333333333,
+      "grad_norm": 0.2690509557723999,
+      "learning_rate": 3e-06,
+      "loss": -0.0359,
+      "policy/loss": 0.03503909707069397,
+      "step": 1216
+    },
+    {
+      "completion_length": 239.17708587646484,
+      "epoch": 0.6490666666666667,
+      "grad_norm": 0.31970566511154175,
+      "learning_rate": 3e-06,
+      "loss": -0.0152,
+      "policy/loss": -0.18568499013781548,
+      "reward": 1.921875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 1.2083333134651184,
+      "step": 1217
+    },
+    {
+      "epoch": 0.6496,
+      "grad_norm": 0.3270527720451355,
+      "learning_rate": 3e-06,
+      "loss": -0.1617,
+      "policy/loss": -0.3786972761154175,
+      "step": 1218
+    },
+    {
+      "epoch": 0.6501333333333333,
+      "grad_norm": 0.36133190989494324,
+      "learning_rate": 3e-06,
+      "loss": -0.0167,
+      "policy/loss": -0.18735924363136292,
+      "step": 1219
+    },
+    {
+      "epoch": 0.6506666666666666,
+      "grad_norm": 0.5127874612808228,
+      "learning_rate": 3e-06,
+      "loss": -0.1663,
+      "policy/loss": -0.3806093633174896,
+      "step": 1220
+    },
+    {
+      "completion_length": 240.77083587646484,
+      "epoch": 0.6512,
+      "grad_norm": 0.25025004148483276,
+      "learning_rate": 3e-06,
+      "loss": 0.0073,
+      "policy/loss": 8.33340187966769e-08,
+      "reward": 1.4375000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 1221
+    },
+    {
+      "epoch": 0.6517333333333334,
+      "grad_norm": 0.1949652135372162,
+      "learning_rate": 3e-06,
+      "loss": 0.0985,
+      "policy/loss": 9.155080959999395e-08,
+      "step": 1222
+    },
+    {
+      "epoch": 0.6522666666666667,
+      "grad_norm": 0.23718786239624023,
+      "learning_rate": 3e-06,
+      "loss": 0.0073,
+      "policy/loss": 8.335337042808533e-08,
+      "step": 1223
+    },
+    {
+      "epoch": 0.6528,
+      "grad_norm": 0.22583569586277008,
+      "learning_rate": 3e-06,
+      "loss": 0.0981,
+      "policy/loss": 9.149147928155799e-08,
+      "step": 1224
+    },
+    {
+      "completion_length": 221.0104217529297,
+      "epoch": 0.6533333333333333,
+      "grad_norm": 0.20737962424755096,
+      "learning_rate": 3e-06,
+      "loss": -0.1005,
+      "policy/loss": -0.4175289645791054,
+      "reward": 1.4843750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.770833358168602,
+      "step": 1225
+    },
+    {
+      "epoch": 0.6538666666666667,
+      "grad_norm": 0.18371179699897766,
+      "learning_rate": 3e-06,
+      "loss": -0.0221,
+      "policy/loss": 0.5971155762672424,
+      "step": 1226
+    },
+    {
+      "epoch": 0.6544,
+      "grad_norm": 0.21121184527873993,
+      "learning_rate": 3e-06,
+      "loss": -0.102,
+      "policy/loss": -0.42161643505096436,
+      "step": 1227
+    },
+    {
+      "epoch": 0.6549333333333334,
+      "grad_norm": 0.1807742416858673,
+      "learning_rate": 3e-06,
+      "loss": -0.0217,
+      "policy/loss": 0.6007712855935097,
+      "step": 1228
+    },
+    {
+      "completion_length": 239.2604217529297,
+      "epoch": 0.6554666666666666,
+      "grad_norm": 0.22973200678825378,
+      "learning_rate": 3e-06,
+      "loss": -0.1805,
+      "policy/loss": 1.2801170967691178e-07,
+      "reward": 1.2552083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.6041666716337204,
+      "step": 1229
+    },
+    {
+      "epoch": 0.656,
+      "grad_norm": 2.2000508308410645,
+      "learning_rate": 3e-06,
+      "loss": -0.0535,
+      "policy/loss": 1.4758767719058596e-07,
+      "step": 1230
+    },
+    {
+      "epoch": 0.6565333333333333,
+      "grad_norm": 0.22820556163787842,
+      "learning_rate": 3e-06,
+      "loss": -0.1809,
+      "policy/loss": 1.2804232696339568e-07,
+      "step": 1231
+    },
+    {
+      "epoch": 0.6570666666666667,
+      "grad_norm": 0.25517377257347107,
+      "learning_rate": 3e-06,
+      "loss": -0.0545,
+      "policy/loss": 1.4783564949993888e-07,
+      "step": 1232
+    },
+    {
+      "completion_length": 219.5416717529297,
+      "epoch": 0.6576,
+      "grad_norm": 0.36168116331100464,
+      "learning_rate": 3e-06,
+      "loss": -0.0914,
+      "policy/loss": -0.06845411222831643,
+      "reward": 1.3020833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 1233
+    },
+    {
+      "epoch": 0.6581333333333333,
+      "grad_norm": 0.264786958694458,
+      "learning_rate": 3e-06,
+      "loss": -0.1382,
+      "policy/loss": 0.026427160673794248,
+      "step": 1234
+    },
+    {
+      "epoch": 0.6586666666666666,
+      "grad_norm": 0.3951287865638733,
+      "learning_rate": 3e-06,
+      "loss": -0.0921,
+      "policy/loss": -0.06961354212273818,
+      "step": 1235
+    },
+    {
+      "epoch": 0.6592,
+      "grad_norm": 0.2685506045818329,
+      "learning_rate": 3e-06,
+      "loss": -0.1398,
+      "policy/loss": 0.024204285518159452,
+      "step": 1236
+    },
+    {
+      "completion_length": 243.90625762939453,
+      "epoch": 0.6597333333333333,
+      "grad_norm": 0.33814504742622375,
+      "learning_rate": 3e-06,
+      "loss": 0.0419,
+      "policy/loss": 8.617629987384134e-08,
+      "reward": 1.4166666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7291666716337204,
+      "step": 1237
+    },
+    {
+      "epoch": 0.6602666666666667,
+      "grad_norm": 0.23845221102237701,
+      "learning_rate": 3e-06,
+      "loss": 0.0009,
+      "policy/loss": 6.468468427556218e-08,
+      "step": 1238
+    },
+    {
+      "epoch": 0.6608,
+      "grad_norm": 1.1938872337341309,
+      "learning_rate": 3e-06,
+      "loss": 0.0435,
+      "policy/loss": 8.609229951161979e-08,
+      "step": 1239
+    },
+    {
+      "epoch": 0.6613333333333333,
+      "grad_norm": 0.37637531757354736,
+      "learning_rate": 3e-06,
+      "loss": 0.0001,
+      "policy/loss": 6.451504219739945e-08,
+      "step": 1240
+    },
+    {
+      "completion_length": 235.75000762939453,
+      "epoch": 0.6618666666666667,
+      "grad_norm": 0.23564980924129486,
+      "learning_rate": 3e-06,
+      "loss": 0.121,
+      "policy/loss": 1.213420652845798e-07,
+      "reward": 1.265625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6614583432674408,
+      "rewards/correctness_reward_func_math": 0.6041666641831398,
+      "step": 1241
+    },
+    {
+      "epoch": 0.6624,
+      "grad_norm": 0.28931373357772827,
+      "learning_rate": 3e-06,
+      "loss": 0.1129,
+      "policy/loss": 1.1646084274730129e-07,
+      "step": 1242
+    },
+    {
+      "epoch": 0.6629333333333334,
+      "grad_norm": 0.2332945168018341,
+      "learning_rate": 3e-06,
+      "loss": 0.1209,
+      "policy/loss": 1.2134752225279044e-07,
+      "step": 1243
+    },
+    {
+      "epoch": 0.6634666666666666,
+      "grad_norm": 0.2716885209083557,
+      "learning_rate": 3e-06,
+      "loss": 0.1124,
+      "policy/loss": 1.1635697916290155e-07,
+      "step": 1244
+    },
+    {
+      "completion_length": 238.96875762939453,
+      "epoch": 0.664,
+      "grad_norm": 0.2588280141353607,
+      "learning_rate": 3e-06,
+      "loss": -0.0111,
+      "policy/loss": 1.3702235079904312e-07,
+      "reward": 1.4479166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 1245
+    },
+    {
+      "epoch": 0.6645333333333333,
+      "grad_norm": 0.23035943508148193,
+      "learning_rate": 3e-06,
+      "loss": 0.0664,
+      "policy/loss": 1.429297853405842e-07,
+      "step": 1246
+    },
+    {
+      "epoch": 0.6650666666666667,
+      "grad_norm": 0.24916957318782806,
+      "learning_rate": 3e-06,
+      "loss": -0.0123,
+      "policy/loss": 1.3694103984107642e-07,
+      "step": 1247
+    },
+    {
+      "epoch": 0.6656,
+      "grad_norm": 0.2316351979970932,
+      "learning_rate": 3e-06,
+      "loss": 0.0653,
+      "policy/loss": 1.428680818094108e-07,
+      "step": 1248
+    },
+    {
+      "completion_length": 244.14583587646484,
+      "epoch": 0.6661333333333334,
+      "grad_norm": 0.2093384712934494,
+      "learning_rate": 3e-06,
+      "loss": -0.0222,
+      "policy/loss": 8.065784484756477e-08,
+      "reward": 1.5,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 1249
+    },
+    {
+      "epoch": 0.6666666666666666,
+      "grad_norm": 0.22133482992649078,
+      "learning_rate": 3e-06,
+      "loss": -0.161,
+      "policy/loss": 1.2810846428124023e-07,
+      "step": 1250
+    },
+    {
+      "epoch": 0.6672,
+      "grad_norm": 0.19444046914577484,
+      "learning_rate": 3e-06,
+      "loss": -0.0228,
+      "policy/loss": 8.06387703278233e-08,
+      "step": 1251
+    },
+    {
+      "epoch": 0.6677333333333333,
+      "grad_norm": 0.22239026427268982,
+      "learning_rate": 3e-06,
+      "loss": -0.1626,
+      "policy/loss": 1.2774022550843256e-07,
+      "step": 1252
+    },
+    {
+      "completion_length": 243.61459350585938,
+      "epoch": 0.6682666666666667,
+      "grad_norm": 0.257695734500885,
+      "learning_rate": 3e-06,
+      "loss": -0.1359,
+      "policy/loss": 8.258661132742873e-08,
+      "reward": 1.4322917461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 1253
+    },
+    {
+      "epoch": 0.6688,
+      "grad_norm": 0.20813192427158356,
+      "learning_rate": 3e-06,
+      "loss": -0.0116,
+      "policy/loss": 9.014194368717199e-08,
+      "step": 1254
+    },
+    {
+      "epoch": 0.6693333333333333,
+      "grad_norm": 0.24760164320468903,
+      "learning_rate": 3e-06,
+      "loss": -0.1361,
+      "policy/loss": 8.243502591653851e-08,
+      "step": 1255
+    },
+    {
+      "epoch": 0.6698666666666667,
+      "grad_norm": 0.20235233008861542,
+      "learning_rate": 3e-06,
+      "loss": -0.0127,
+      "policy/loss": 9.010805612774675e-08,
+      "step": 1256
+    },
+    {
+      "completion_length": 240.90625762939453,
+      "epoch": 0.6704,
+      "grad_norm": 0.2369832992553711,
+      "learning_rate": 3e-06,
+      "loss": 0.0825,
+      "policy/loss": -0.16280141325080422,
+      "reward": 1.3020833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6770833134651184,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 1257
+    },
+    {
+      "epoch": 0.6709333333333334,
+      "grad_norm": 0.2170265167951584,
+      "learning_rate": 3e-06,
+      "loss": -0.1997,
+      "policy/loss": -0.11831129504405169,
+      "step": 1258
+    },
+    {
+      "epoch": 0.6714666666666667,
+      "grad_norm": 0.2350655347108841,
+      "learning_rate": 3e-06,
+      "loss": 0.0814,
+      "policy/loss": -0.16158808614273923,
+      "step": 1259
+    },
+    {
+      "epoch": 0.672,
+      "grad_norm": 0.22838819026947021,
+      "learning_rate": 3e-06,
+      "loss": -0.201,
+      "policy/loss": -0.11969702852812425,
+      "step": 1260
+    },
+    {
+      "completion_length": 231.625,
+      "epoch": 0.6725333333333333,
+      "grad_norm": 0.19917157292366028,
+      "learning_rate": 3e-06,
+      "loss": -0.1258,
+      "policy/loss": -0.026889652013778687,
+      "reward": 1.3229166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6041666567325592,
+      "step": 1261
+    },
+    {
+      "epoch": 0.6730666666666667,
+      "grad_norm": 0.18916988372802734,
+      "learning_rate": 3e-06,
+      "loss": -0.0014,
+      "policy/loss": 0.02300172857940197,
+      "step": 1262
+    },
+    {
+      "epoch": 0.6736,
+      "grad_norm": 0.15756526589393616,
+      "learning_rate": 3e-06,
+      "loss": -0.1268,
+      "policy/loss": -0.026415277272462845,
+      "step": 1263
+    },
+    {
+      "epoch": 0.6741333333333334,
+      "grad_norm": 0.18776549398899078,
+      "learning_rate": 3e-06,
+      "loss": -0.0012,
+      "policy/loss": 0.02243754267692566,
+      "step": 1264
+    },
+    {
+      "completion_length": 237.8229217529297,
+      "epoch": 0.6746666666666666,
+      "grad_norm": 0.22136497497558594,
+      "learning_rate": 3e-06,
+      "loss": -0.1734,
+      "policy/loss": 0.01433611923635425,
+      "reward": 1.4375000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 1265
+    },
+    {
+      "epoch": 0.6752,
+      "grad_norm": 0.20838794112205505,
+      "learning_rate": 3e-06,
+      "loss": 0.0912,
+      "policy/loss": 0.1334379101399179,
+      "step": 1266
+    },
+    {
+      "epoch": 0.6757333333333333,
+      "grad_norm": 0.22942809760570526,
+      "learning_rate": 3e-06,
+      "loss": -0.1742,
+      "policy/loss": 0.012608558913115075,
+      "step": 1267
+    },
+    {
+      "epoch": 0.6762666666666667,
+      "grad_norm": 0.2137482911348343,
+      "learning_rate": 3e-06,
+      "loss": 0.0898,
+      "policy/loss": 0.12950892093105892,
+      "step": 1268
+    },
+    {
+      "completion_length": 231.21875,
+      "epoch": 0.6768,
+      "grad_norm": 0.24802635610103607,
+      "learning_rate": 3e-06,
+      "loss": 0.0005,
+      "policy/loss": 0.033243087809797345,
+      "reward": 1.4322916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 1269
+    },
+    {
+      "epoch": 0.6773333333333333,
+      "grad_norm": 0.26831158995628357,
+      "learning_rate": 3e-06,
+      "loss": 0.0712,
+      "policy/loss": 0.6473587728580661,
+      "step": 1270
+    },
+    {
+      "epoch": 0.6778666666666666,
+      "grad_norm": 0.24706198275089264,
+      "learning_rate": 3e-06,
+      "loss": -0.0006,
+      "policy/loss": 0.032725119742924136,
+      "step": 1271
+    },
+    {
+      "epoch": 0.6784,
+      "grad_norm": 0.27716827392578125,
+      "learning_rate": 3e-06,
+      "loss": 0.0704,
+      "policy/loss": 0.6507583237794847,
+      "step": 1272
+    },
+    {
+      "completion_length": 234.20833587646484,
+      "epoch": 0.6789333333333334,
+      "grad_norm": 0.25065892934799194,
+      "learning_rate": 3e-06,
+      "loss": -0.0942,
+      "policy/loss": -0.19994761205243172,
+      "reward": 1.1250000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333432674408,
+      "rewards/correctness_reward_func_math": 0.4166666716337204,
+      "step": 1273
+    },
+    {
+      "epoch": 0.6794666666666667,
+      "grad_norm": 0.2323308140039444,
+      "learning_rate": 3e-06,
+      "loss": 0.0623,
+      "policy/loss": 0.03457993551474736,
+      "step": 1274
+    },
+    {
+      "epoch": 0.68,
+      "grad_norm": 0.2615307867527008,
+      "learning_rate": 3e-06,
+      "loss": -0.0955,
+      "policy/loss": -0.20243224653521086,
+      "step": 1275
+    },
+    {
+      "epoch": 0.6805333333333333,
+      "grad_norm": 0.2377048283815384,
+      "learning_rate": 3e-06,
+      "loss": 0.0601,
+      "policy/loss": 0.030430561884091212,
+      "step": 1276
+    },
+    {
+      "completion_length": 239.05208587646484,
+      "epoch": 0.6810666666666667,
+      "grad_norm": 0.21531617641448975,
+      "learning_rate": 3e-06,
+      "loss": 0.0511,
+      "policy/loss": -0.13643565165438432,
+      "reward": 1.4635416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 1277
+    },
+    {
+      "epoch": 0.6816,
+      "grad_norm": 0.24599704146385193,
+      "learning_rate": 3e-06,
+      "loss": 0.0082,
+      "policy/loss": -0.06215125919058195,
+      "step": 1278
+    },
+    {
+      "epoch": 0.6821333333333334,
+      "grad_norm": 0.20688703656196594,
+      "learning_rate": 3e-06,
+      "loss": 0.0498,
+      "policy/loss": -0.13671013097144424,
+      "step": 1279
+    },
+    {
+      "epoch": 0.6826666666666666,
+      "grad_norm": 0.2808336317539215,
+      "learning_rate": 3e-06,
+      "loss": 0.0087,
+      "policy/loss": -0.06238170559826628,
+      "step": 1280
+    },
+    {
+      "completion_length": 237.7604217529297,
+      "epoch": 0.6832,
+      "grad_norm": 0.2322651743888855,
+      "learning_rate": 3e-06,
+      "loss": 0.0361,
+      "policy/loss": 1.3794327990979127e-07,
+      "reward": 1.3854167461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.7291666865348816,
+      "step": 1281
+    },
+    {
+      "epoch": 0.6837333333333333,
+      "grad_norm": 0.2823871970176697,
+      "learning_rate": 3e-06,
+      "loss": -0.0153,
+      "policy/loss": 1.1375880148989381e-07,
+      "step": 1282
+    },
+    {
+      "epoch": 0.6842666666666667,
+      "grad_norm": 0.22885338962078094,
+      "learning_rate": 3e-06,
+      "loss": 0.035,
+      "policy/loss": 1.3797579434537965e-07,
+      "step": 1283
+    },
+    {
+      "epoch": 0.6848,
+      "grad_norm": 0.29989829659461975,
+      "learning_rate": 3e-06,
+      "loss": -0.0151,
+      "policy/loss": 1.134412901393489e-07,
+      "step": 1284
+    },
+    {
+      "completion_length": 237.90625762939453,
+      "epoch": 0.6853333333333333,
+      "grad_norm": 0.20473505556583405,
+      "learning_rate": 3e-06,
+      "loss": -0.037,
+      "policy/loss": 0.3091107549303054,
+      "reward": 1.0729166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.3541666716337204,
+      "step": 1285
+    },
+    {
+      "epoch": 0.6858666666666666,
+      "grad_norm": 0.19564664363861084,
+      "learning_rate": 3e-06,
+      "loss": -0.0423,
+      "policy/loss": -0.0947610711594109,
+      "step": 1286
+    },
+    {
+      "epoch": 0.6864,
+      "grad_norm": 0.18183305859565735,
+      "learning_rate": 3e-06,
+      "loss": -0.0384,
+      "policy/loss": 0.3056724610539874,
+      "step": 1287
+    },
+    {
+      "epoch": 0.6869333333333333,
+      "grad_norm": 0.19729477167129517,
+      "learning_rate": 3e-06,
+      "loss": -0.0433,
+      "policy/loss": -0.09612890836390875,
+      "step": 1288
+    },
+    {
+      "completion_length": 235.09375762939453,
+      "epoch": 0.6874666666666667,
+      "grad_norm": 0.22296780347824097,
+      "learning_rate": 3e-06,
+      "loss": 0.0792,
+      "policy/loss": -0.09363659343631525,
+      "reward": 1.3750000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.6250000149011612,
+      "step": 1289
+    },
+    {
+      "epoch": 0.688,
+      "grad_norm": 0.29955849051475525,
+      "learning_rate": 3e-06,
+      "loss": -0.1036,
+      "policy/loss": 0.06025634669468616,
+      "step": 1290
+    },
+    {
+      "epoch": 0.6885333333333333,
+      "grad_norm": 0.22894041240215302,
+      "learning_rate": 3e-06,
+      "loss": 0.0787,
+      "policy/loss": -0.09559420378138839,
+      "step": 1291
+    },
+    {
+      "epoch": 0.6890666666666667,
+      "grad_norm": 0.2994455397129059,
+      "learning_rate": 3e-06,
+      "loss": -0.1045,
+      "policy/loss": 0.06150977276784175,
+      "step": 1292
+    },
+    {
+      "completion_length": 244.31250762939453,
+      "epoch": 0.6896,
+      "grad_norm": 0.19555409252643585,
+      "learning_rate": 3e-06,
+      "loss": -0.0545,
+      "policy/loss": 0.0002590865898319805,
+      "reward": 1.1250000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.375,
+      "step": 1293
+    },
+    {
+      "epoch": 0.6901333333333334,
+      "grad_norm": 0.25190383195877075,
+      "learning_rate": 3e-06,
+      "loss": 0.0601,
+      "policy/loss": 0.1565284462733949,
+      "step": 1294
+    },
+    {
+      "epoch": 0.6906666666666667,
+      "grad_norm": 0.19037778675556183,
+      "learning_rate": 3e-06,
+      "loss": -0.0551,
+      "policy/loss": -0.0006714672891625639,
+      "step": 1295
+    },
+    {
+      "epoch": 0.6912,
+      "grad_norm": 0.27696990966796875,
+      "learning_rate": 3e-06,
+      "loss": 0.059,
+      "policy/loss": 0.1598127068618389,
+      "step": 1296
+    },
+    {
+      "completion_length": 233.5729217529297,
+      "epoch": 0.6917333333333333,
+      "grad_norm": 0.16817224025726318,
+      "learning_rate": 3e-06,
+      "loss": 0.1368,
+      "policy/loss": 0.09781777858734131,
+      "reward": 1.1458333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.4583333432674408,
+      "step": 1297
+    },
+    {
+      "epoch": 0.6922666666666667,
+      "grad_norm": 0.16950735449790955,
+      "learning_rate": 3e-06,
+      "loss": 0.0591,
+      "policy/loss": 0.30839312076568604,
+      "step": 1298
+    },
+    {
+      "epoch": 0.6928,
+      "grad_norm": 0.1732439547777176,
+      "learning_rate": 3e-06,
+      "loss": 0.1356,
+      "policy/loss": 0.09307282418012619,
+      "step": 1299
+    },
+    {
+      "epoch": 0.6933333333333334,
+      "grad_norm": 0.1635216474533081,
+      "learning_rate": 3e-06,
+      "loss": 0.0588,
+      "policy/loss": 0.3095167875289917,
+      "step": 1300
+    },
+    {
+      "completion_length": 235.95833587646484,
+      "epoch": 0.6938666666666666,
+      "grad_norm": 0.2798515856266022,
+      "learning_rate": 3e-06,
+      "loss": -0.0085,
+      "policy/loss": -0.03713326156139374,
+      "reward": 1.5312500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.875,
+      "step": 1301
+    },
+    {
+      "epoch": 0.6944,
+      "grad_norm": 0.2037225365638733,
+      "learning_rate": 3e-06,
+      "loss": -0.0326,
+      "policy/loss": 0.006525305099785328,
+      "step": 1302
+    },
+    {
+      "epoch": 0.6949333333333333,
+      "grad_norm": 0.32651180028915405,
+      "learning_rate": 3e-06,
+      "loss": -0.0093,
+      "policy/loss": -0.0398266538977623,
+      "step": 1303
+    },
+    {
+      "epoch": 0.6954666666666667,
+      "grad_norm": 0.2162570357322693,
+      "learning_rate": 3e-06,
+      "loss": -0.0334,
+      "policy/loss": 0.004861287772655487,
+      "step": 1304
+    },
+    {
+      "completion_length": 235.2604217529297,
+      "epoch": 0.696,
+      "grad_norm": 0.24084177613258362,
+      "learning_rate": 3e-06,
+      "loss": 0.0385,
+      "policy/loss": 1.0615186596396597e-07,
+      "reward": 1.1875000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6666666567325592,
+      "rewards/correctness_reward_func_math": 0.5208333432674408,
+      "step": 1305
+    },
+    {
+      "epoch": 0.6965333333333333,
+      "grad_norm": 0.2610556185245514,
+      "learning_rate": 3e-06,
+      "loss": 0.1521,
+      "policy/loss": 9.152699931291863e-08,
+      "step": 1306
+    },
+    {
+      "epoch": 0.6970666666666666,
+      "grad_norm": 0.23485717177391052,
+      "learning_rate": 3e-06,
+      "loss": 0.0377,
+      "policy/loss": 1.0609358547242209e-07,
+      "step": 1307
+    },
+    {
+      "epoch": 0.6976,
+      "grad_norm": 0.2557401955127716,
+      "learning_rate": 3e-06,
+      "loss": 0.1505,
+      "policy/loss": 9.154014612988703e-08,
+      "step": 1308
+    },
+    {
+      "completion_length": 237.95833587646484,
+      "epoch": 0.6981333333333334,
+      "grad_norm": 0.18872955441474915,
+      "learning_rate": 3e-06,
+      "loss": 0.0461,
+      "policy/loss": -0.004711002973621703,
+      "reward": 1.4114583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.6666666567325592,
+      "step": 1309
+    },
+    {
+      "epoch": 0.6986666666666667,
+      "grad_norm": 0.20233504474163055,
+      "learning_rate": 3e-06,
+      "loss": -0.0103,
+      "policy/loss": 0.3098062713539811,
+      "step": 1310
+    },
+    {
+      "epoch": 0.6992,
+      "grad_norm": 0.18261869251728058,
+      "learning_rate": 3e-06,
+      "loss": 0.0454,
+      "policy/loss": -0.004411332347743269,
+      "step": 1311
+    },
+    {
+      "epoch": 0.6997333333333333,
+      "grad_norm": 0.19652266800403595,
+      "learning_rate": 3e-06,
+      "loss": -0.0103,
+      "policy/loss": 0.30925156072996884,
+      "step": 1312
+    },
+    {
+      "completion_length": 234.71875,
+      "epoch": 0.7002666666666667,
+      "grad_norm": 0.24839478731155396,
+      "learning_rate": 3e-06,
+      "loss": -0.1178,
+      "policy/loss": -0.12987476587295532,
+      "reward": 1.6250000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.9374999701976776,
+      "step": 1313
+    },
+    {
+      "epoch": 0.7008,
+      "grad_norm": 0.31159183382987976,
+      "learning_rate": 3e-06,
+      "loss": -0.0664,
+      "policy/loss": 0.031507208943367004,
+      "step": 1314
+    },
+    {
+      "epoch": 0.7013333333333334,
+      "grad_norm": 0.2360585778951645,
+      "learning_rate": 3e-06,
+      "loss": -0.1184,
+      "policy/loss": -0.12841834127902985,
+      "step": 1315
+    },
+    {
+      "epoch": 0.7018666666666666,
+      "grad_norm": 0.27359697222709656,
+      "learning_rate": 3e-06,
+      "loss": -0.0673,
+      "policy/loss": 0.03189562261104584,
+      "step": 1316
+    },
+    {
+      "completion_length": 231.8541717529297,
+      "epoch": 0.7024,
+      "grad_norm": 0.30645495653152466,
+      "learning_rate": 3e-06,
+      "loss": 0.1156,
+      "policy/loss": 0.11942961812019348,
+      "reward": 1.5416666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333432674408,
+      "rewards/correctness_reward_func_math": 0.8333333432674408,
+      "step": 1317
+    },
+    {
+      "epoch": 0.7029333333333333,
+      "grad_norm": 0.23117956519126892,
+      "learning_rate": 3e-06,
+      "loss": -0.0162,
+      "policy/loss": -0.306593120098114,
+      "step": 1318
+    },
+    {
+      "epoch": 0.7034666666666667,
+      "grad_norm": 0.2829439342021942,
+      "learning_rate": 3e-06,
+      "loss": 0.1161,
+      "policy/loss": 0.1201077550649643,
+      "step": 1319
+    },
+    {
+      "epoch": 0.704,
+      "grad_norm": 0.2289082407951355,
+      "learning_rate": 3e-06,
+      "loss": -0.0172,
+      "policy/loss": -0.30963608622550964,
+      "step": 1320
+    },
+    {
+      "completion_length": 234.5104217529297,
+      "epoch": 0.7045333333333333,
+      "grad_norm": 0.23192495107650757,
+      "learning_rate": 3e-06,
+      "loss": 0.1307,
+      "policy/loss": 1.0228659874655932e-07,
+      "reward": 1.2916666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.6666666567325592,
+      "step": 1321
+    },
+    {
+      "epoch": 0.7050666666666666,
+      "grad_norm": 0.25179585814476013,
+      "learning_rate": 3e-06,
+      "loss": -0.0592,
+      "policy/loss": 8.870035372865459e-08,
+      "step": 1322
+    },
+    {
+      "epoch": 0.7056,
+      "grad_norm": 0.2280527651309967,
+      "learning_rate": 3e-06,
+      "loss": 0.1297,
+      "policy/loss": 1.0214061418878373e-07,
+      "step": 1323
+    },
+    {
+      "epoch": 0.7061333333333333,
+      "grad_norm": 0.24429155886173248,
+      "learning_rate": 3e-06,
+      "loss": -0.0602,
+      "policy/loss": 8.857381317284307e-08,
+      "step": 1324
+    },
+    {
+      "completion_length": 239.92708587646484,
+      "epoch": 0.7066666666666667,
+      "grad_norm": 0.18611198663711548,
+      "learning_rate": 3e-06,
+      "loss": 0.0179,
+      "policy/loss": 0.11726571841975897,
+      "reward": 1.3020833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.6458333134651184,
+      "step": 1325
+    },
+    {
+      "epoch": 0.7072,
+      "grad_norm": 0.16426555812358856,
+      "learning_rate": 3e-06,
+      "loss": 0.0566,
+      "policy/loss": -0.2226199519448251,
+      "step": 1326
+    },
+    {
+      "epoch": 0.7077333333333333,
+      "grad_norm": 0.1663249433040619,
+      "learning_rate": 3e-06,
+      "loss": 0.018,
+      "policy/loss": 0.11872906473251987,
+      "step": 1327
+    },
+    {
+      "epoch": 0.7082666666666667,
+      "grad_norm": 0.16871310770511627,
+      "learning_rate": 3e-06,
+      "loss": 0.0556,
+      "policy/loss": -0.22276966404001897,
+      "step": 1328
+    },
+    {
+      "completion_length": 240.30209350585938,
+      "epoch": 0.7088,
+      "grad_norm": 0.18763527274131775,
+      "learning_rate": 3e-06,
+      "loss": -0.006,
+      "policy/loss": 0.5290040820837021,
+      "reward": 1.4010416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 1329
+    },
+    {
+      "epoch": 0.7093333333333334,
+      "grad_norm": 0.190482497215271,
+      "learning_rate": 3e-06,
+      "loss": 0.0036,
+      "policy/loss": 0.547521784901619,
+      "step": 1330
+    },
+    {
+      "epoch": 0.7098666666666666,
+      "grad_norm": 0.17721787095069885,
+      "learning_rate": 3e-06,
+      "loss": -0.007,
+      "policy/loss": 0.5264172405004501,
+      "step": 1331
+    },
+    {
+      "epoch": 0.7104,
+      "grad_norm": 0.19178873300552368,
+      "learning_rate": 3e-06,
+      "loss": 0.002,
+      "policy/loss": 0.5434936694800854,
+      "step": 1332
+    },
+    {
+      "completion_length": 232.0416717529297,
+      "epoch": 0.7109333333333333,
+      "grad_norm": 0.24506355822086334,
+      "learning_rate": 3e-06,
+      "loss": 0.0224,
+      "policy/loss": -0.01125900844476746,
+      "reward": 1.6614583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.9166666567325592,
+      "step": 1333
+    },
+    {
+      "epoch": 0.7114666666666667,
+      "grad_norm": 0.21428057551383972,
+      "learning_rate": 3e-06,
+      "loss": 0.0188,
+      "policy/loss": 0.06237173860870371,
+      "step": 1334
+    },
+    {
+      "epoch": 0.712,
+      "grad_norm": 0.23723039031028748,
+      "learning_rate": 3e-06,
+      "loss": 0.0214,
+      "policy/loss": -0.011434599071023399,
+      "step": 1335
+    },
+    {
+      "epoch": 0.7125333333333334,
+      "grad_norm": 0.20599493384361267,
+      "learning_rate": 3e-06,
+      "loss": 0.0182,
+      "policy/loss": 0.06209615638374544,
+      "step": 1336
+    },
+    {
+      "completion_length": 245.05208587646484,
+      "epoch": 0.7130666666666666,
+      "grad_norm": 0.2417358160018921,
+      "learning_rate": 3e-06,
+      "loss": -0.1249,
+      "policy/loss": -0.3395807258784771,
+      "reward": 1.4531250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.770833358168602,
+      "step": 1337
+    },
+    {
+      "epoch": 0.7136,
+      "grad_norm": 0.24642620980739594,
+      "learning_rate": 3e-06,
+      "loss": 0.0851,
+      "policy/loss": -0.12109694257378578,
+      "step": 1338
+    },
+    {
+      "epoch": 0.7141333333333333,
+      "grad_norm": 0.2197198122739792,
+      "learning_rate": 3e-06,
+      "loss": -0.1245,
+      "policy/loss": -0.33533033542335033,
+      "step": 1339
+    },
+    {
+      "epoch": 0.7146666666666667,
+      "grad_norm": 0.2438141107559204,
+      "learning_rate": 3e-06,
+      "loss": 0.0849,
+      "policy/loss": -0.11414780467748642,
+      "step": 1340
+    },
+    {
+      "completion_length": 242.20833587646484,
+      "epoch": 0.7152,
+      "grad_norm": 0.2818518280982971,
+      "learning_rate": 3e-06,
+      "loss": -0.1646,
+      "policy/loss": 0.7489531934261322,
+      "reward": 1.1822916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.4375,
+      "step": 1341
+    },
+    {
+      "epoch": 0.7157333333333333,
+      "grad_norm": 0.2628598213195801,
+      "learning_rate": 3e-06,
+      "loss": -0.0832,
+      "policy/loss": 0.4064641147851944,
+      "step": 1342
+    },
+    {
+      "epoch": 0.7162666666666667,
+      "grad_norm": 0.24663567543029785,
+      "learning_rate": 3e-06,
+      "loss": -0.1651,
+      "policy/loss": 0.7486852705478668,
+      "step": 1343
+    },
+    {
+      "epoch": 0.7168,
+      "grad_norm": 0.25506865978240967,
+      "learning_rate": 3e-06,
+      "loss": -0.0842,
+      "policy/loss": 0.40227891504764557,
+      "step": 1344
+    },
+    {
+      "completion_length": 242.1354217529297,
+      "epoch": 0.7173333333333334,
+      "grad_norm": 0.23433412611484528,
+      "learning_rate": 3e-06,
+      "loss": -0.1162,
+      "policy/loss": 1.6689045878592879e-07,
+      "reward": 1.2083333432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.5208333283662796,
+      "step": 1345
+    },
+    {
+      "epoch": 0.7178666666666667,
+      "grad_norm": 0.24479006230831146,
+      "learning_rate": 3e-06,
+      "loss": 0.1465,
+      "policy/loss": 1.6882115971839085e-07,
+      "step": 1346
+    },
+    {
+      "epoch": 0.7184,
+      "grad_norm": 0.25713297724723816,
+      "learning_rate": 3e-06,
+      "loss": -0.1188,
+      "policy/loss": 1.6690618309667116e-07,
+      "step": 1347
+    },
+    {
+      "epoch": 0.7189333333333333,
+      "grad_norm": 0.294807106256485,
+      "learning_rate": 3e-06,
+      "loss": 0.1448,
+      "policy/loss": 1.6878681918797156e-07,
+      "step": 1348
+    },
+    {
+      "completion_length": 224.42709350585938,
+      "epoch": 0.7194666666666667,
+      "grad_norm": 0.2681988477706909,
+      "learning_rate": 3e-06,
+      "loss": 0.031,
+      "policy/loss": 0.034621529530880935,
+      "reward": 1.5989583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6197916567325592,
+      "rewards/correctness_reward_func_math": 0.9791666865348816,
+      "step": 1349
+    },
+    {
+      "epoch": 0.72,
+      "grad_norm": 0.2506592869758606,
+      "learning_rate": 3e-06,
+      "loss": -0.0215,
+      "policy/loss": -0.03962410377371128,
+      "step": 1350
+    },
+    {
+      "epoch": 0.7205333333333334,
+      "grad_norm": 0.2618780732154846,
+      "learning_rate": 3e-06,
+      "loss": 0.0306,
+      "policy/loss": 0.035045799110470455,
+      "step": 1351
+    },
+    {
+      "epoch": 0.7210666666666666,
+      "grad_norm": 0.24865734577178955,
+      "learning_rate": 3e-06,
+      "loss": -0.022,
+      "policy/loss": -0.04110825932862028,
+      "step": 1352
+    },
+    {
+      "completion_length": 237.1354217529297,
+      "epoch": 0.7216,
+      "grad_norm": 0.407006174325943,
+      "learning_rate": 3e-06,
+      "loss": -0.0991,
+      "policy/loss": 7.093075282682548e-08,
+      "reward": 1.4583333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7708333432674408,
+      "step": 1353
+    },
+    {
+      "epoch": 0.7221333333333333,
+      "grad_norm": 0.259961873292923,
+      "learning_rate": 3e-06,
+      "loss": -0.0527,
+      "policy/loss": 7.975514471070255e-08,
+      "step": 1354
+    },
+    {
+      "epoch": 0.7226666666666667,
+      "grad_norm": 0.4034102261066437,
+      "learning_rate": 3e-06,
+      "loss": -0.1,
+      "policy/loss": 7.090839737600163e-08,
+      "step": 1355
+    },
+    {
+      "epoch": 0.7232,
+      "grad_norm": 0.26698872447013855,
+      "learning_rate": 3e-06,
+      "loss": -0.0552,
+      "policy/loss": 7.960335146606212e-08,
+      "step": 1356
+    },
+    {
+      "completion_length": 237.95833587646484,
+      "epoch": 0.7237333333333333,
+      "grad_norm": 0.21923615038394928,
+      "learning_rate": 3e-06,
+      "loss": -0.3357,
+      "policy/loss": 6.895096404946344e-08,
+      "reward": 1.2291666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.5416666716337204,
+      "step": 1357
+    },
+    {
+      "epoch": 0.7242666666666666,
+      "grad_norm": 0.2137162834405899,
+      "learning_rate": 3e-06,
+      "loss": -0.1439,
+      "policy/loss": 6.885155201530324e-08,
+      "step": 1358
+    },
+    {
+      "epoch": 0.7248,
+      "grad_norm": 0.22243452072143555,
+      "learning_rate": 3e-06,
+      "loss": -0.3358,
+      "policy/loss": 6.885195347194895e-08,
+      "step": 1359
+    },
+    {
+      "epoch": 0.7253333333333334,
+      "grad_norm": 0.21413688361644745,
+      "learning_rate": 3e-06,
+      "loss": -0.1448,
+      "policy/loss": 6.884955539021576e-08,
+      "step": 1360
+    },
+    {
+      "completion_length": 242.56250762939453,
+      "epoch": 0.7258666666666667,
+      "grad_norm": 0.2051650732755661,
+      "learning_rate": 3e-06,
+      "loss": -0.0096,
+      "policy/loss": 1.1723211912340048e-07,
+      "reward": 1.3385417461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 1361
+    },
+    {
+      "epoch": 0.7264,
+      "grad_norm": 0.29902809858322144,
+      "learning_rate": 3e-06,
+      "loss": 0.0827,
+      "policy/loss": 1.237216871174951e-07,
+      "step": 1362
+    },
+    {
+      "epoch": 0.7269333333333333,
+      "grad_norm": 0.2069276124238968,
+      "learning_rate": 3e-06,
+      "loss": -0.0097,
+      "policy/loss": 1.1711811609416145e-07,
+      "step": 1363
+    },
+    {
+      "epoch": 0.7274666666666667,
+      "grad_norm": 0.43395304679870605,
+      "learning_rate": 3e-06,
+      "loss": 0.0819,
+      "policy/loss": 1.2356718315231774e-07,
+      "step": 1364
+    },
+    {
+      "completion_length": 239.05208587646484,
+      "epoch": 0.728,
+      "grad_norm": 0.1799820512533188,
+      "learning_rate": 3e-06,
+      "loss": -0.1646,
+      "policy/loss": 0.007698192402379789,
+      "reward": 1.2812500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.5625,
+      "step": 1365
+    },
+    {
+      "epoch": 0.7285333333333334,
+      "grad_norm": 0.1810152679681778,
+      "learning_rate": 3e-06,
+      "loss": -0.048,
+      "policy/loss": -0.48001185867695995,
+      "step": 1366
+    },
+    {
+      "epoch": 0.7290666666666666,
+      "grad_norm": 0.1785617470741272,
+      "learning_rate": 3e-06,
+      "loss": -0.1658,
+      "policy/loss": 0.002044205195943505,
+      "step": 1367
+    },
+    {
+      "epoch": 0.7296,
+      "grad_norm": 0.19020015001296997,
+      "learning_rate": 3e-06,
+      "loss": -0.0492,
+      "policy/loss": -0.47925735352527354,
+      "step": 1368
+    },
+    {
+      "completion_length": 222.71875762939453,
+      "epoch": 0.7301333333333333,
+      "grad_norm": 0.18686383962631226,
+      "learning_rate": 3e-06,
+      "loss": -0.1021,
+      "policy/loss": -1.1647093137174966,
+      "reward": 1.4062500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 1369
+    },
+    {
+      "epoch": 0.7306666666666667,
+      "grad_norm": 0.23867155611515045,
+      "learning_rate": 3e-06,
+      "loss": -0.118,
+      "policy/loss": -0.8840371171209842,
+      "step": 1370
+    },
+    {
+      "epoch": 0.7312,
+      "grad_norm": 0.19078131020069122,
+      "learning_rate": 3e-06,
+      "loss": -0.1038,
+      "policy/loss": -1.1741596301386608,
+      "step": 1371
+    },
+    {
+      "epoch": 0.7317333333333333,
+      "grad_norm": 0.23707839846611023,
+      "learning_rate": 3e-06,
+      "loss": -0.1191,
+      "policy/loss": -0.8864857951772205,
+      "step": 1372
+    },
+    {
+      "completion_length": 237.0625,
+      "epoch": 0.7322666666666666,
+      "grad_norm": 0.39052626490592957,
+      "learning_rate": 3e-06,
+      "loss": -0.0562,
+      "policy/loss": 7.16312964499366e-08,
+      "reward": 1.3645833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.7083333432674408,
+      "step": 1373
+    },
+    {
+      "epoch": 0.7328,
+      "grad_norm": 0.2272036224603653,
+      "learning_rate": 3e-06,
+      "loss": -0.1503,
+      "policy/loss": 6.609136882218536e-08,
+      "step": 1374
+    },
+    {
+      "epoch": 0.7333333333333333,
+      "grad_norm": 0.2693171799182892,
+      "learning_rate": 3e-06,
+      "loss": -0.0572,
+      "policy/loss": 7.164302573414716e-08,
+      "step": 1375
+    },
+    {
+      "epoch": 0.7338666666666667,
+      "grad_norm": 0.2162552773952484,
+      "learning_rate": 3e-06,
+      "loss": -0.1509,
+      "policy/loss": 6.618161840776793e-08,
+      "step": 1376
+    },
+    {
+      "completion_length": 227.67708587646484,
+      "epoch": 0.7344,
+      "grad_norm": 0.3007422089576721,
+      "learning_rate": 3e-06,
+      "loss": 0.045,
+      "policy/loss": -0.16758218951440895,
+      "reward": 1.2552083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.5416666716337204,
+      "step": 1377
+    },
+    {
+      "epoch": 0.7349333333333333,
+      "grad_norm": 0.3039637804031372,
+      "learning_rate": 3e-06,
+      "loss": -0.0922,
+      "policy/loss": -0.1837413440409179,
+      "step": 1378
+    },
+    {
+      "epoch": 0.7354666666666667,
+      "grad_norm": 0.27520376443862915,
+      "learning_rate": 3e-06,
+      "loss": 0.0435,
+      "policy/loss": -0.16661190041596186,
+      "step": 1379
+    },
+    {
+      "epoch": 0.736,
+      "grad_norm": 0.3091704547405243,
+      "learning_rate": 3e-06,
+      "loss": -0.0942,
+      "policy/loss": -0.18473101957555826,
+      "step": 1380
+    },
+    {
+      "completion_length": 235.46875,
+      "epoch": 0.7365333333333334,
+      "grad_norm": 0.277332603931427,
+      "learning_rate": 3e-06,
+      "loss": -0.0296,
+      "policy/loss": 3.077054344657881e-08,
+      "reward": 1.1041666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.4166666716337204,
+      "step": 1381
+    },
+    {
+      "epoch": 0.7370666666666666,
+      "grad_norm": 0.34084808826446533,
+      "learning_rate": 3e-06,
+      "loss": -0.047,
+      "policy/loss": 4.2860524018806245e-08,
+      "step": 1382
+    },
+    {
+      "epoch": 0.7376,
+      "grad_norm": 0.2715982496738434,
+      "learning_rate": 3e-06,
+      "loss": -0.0298,
+      "policy/loss": 3.077168031495603e-08,
+      "step": 1383
+    },
+    {
+      "epoch": 0.7381333333333333,
+      "grad_norm": 0.32423168420791626,
+      "learning_rate": 3e-06,
+      "loss": -0.0471,
+      "policy/loss": 4.281702103980933e-08,
+      "step": 1384
+    },
+    {
+      "completion_length": 222.1979217529297,
+      "epoch": 0.7386666666666667,
+      "grad_norm": 0.22037512063980103,
+      "learning_rate": 3e-06,
+      "loss": -0.0394,
+      "policy/loss": 0.06564132124185562,
+      "reward": 1.5052083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 1385
+    },
+    {
+      "epoch": 0.7392,
+      "grad_norm": 0.20341695845127106,
+      "learning_rate": 3e-06,
+      "loss": -0.0609,
+      "policy/loss": -0.04244820401072502,
+      "step": 1386
+    },
+    {
+      "epoch": 0.7397333333333334,
+      "grad_norm": 0.2209702581167221,
+      "learning_rate": 3e-06,
+      "loss": -0.0409,
+      "policy/loss": 0.0660535916686058,
+      "step": 1387
+    },
+    {
+      "epoch": 0.7402666666666666,
+      "grad_norm": 0.18818847835063934,
+      "learning_rate": 3e-06,
+      "loss": -0.0619,
+      "policy/loss": -0.041688304394483566,
+      "step": 1388
+    },
+    {
+      "completion_length": 230.81250762939453,
+      "epoch": 0.7408,
+      "grad_norm": 0.24324804544448853,
+      "learning_rate": 3e-06,
+      "loss": -0.0675,
+      "policy/loss": -0.22118137141142746,
+      "reward": 1.2343750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5885416567325592,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 1389
+    },
+    {
+      "epoch": 0.7413333333333333,
+      "grad_norm": 0.3332115411758423,
+      "learning_rate": 3e-06,
+      "loss": 0.0704,
+      "policy/loss": 0.09895377891604085,
+      "step": 1390
+    },
+    {
+      "epoch": 0.7418666666666667,
+      "grad_norm": 0.22708697617053986,
+      "learning_rate": 3e-06,
+      "loss": -0.0688,
+      "policy/loss": -0.22229042010639155,
+      "step": 1391
+    },
+    {
+      "epoch": 0.7424,
+      "grad_norm": 0.32536372542381287,
+      "learning_rate": 3e-06,
+      "loss": 0.0696,
+      "policy/loss": 0.10089464769977852,
+      "step": 1392
+    },
+    {
+      "completion_length": 244.1041717529297,
+      "epoch": 0.7429333333333333,
+      "grad_norm": 0.2250286489725113,
+      "learning_rate": 3e-06,
+      "loss": -0.0291,
+      "policy/loss": 0.10210350602821805,
+      "reward": 1.2604166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.5416666567325592,
+      "step": 1393
+    },
+    {
+      "epoch": 0.7434666666666667,
+      "grad_norm": 0.2957366406917572,
+      "learning_rate": 3e-06,
+      "loss": -0.1244,
+      "policy/loss": 0.04385423027639135,
+      "step": 1394
+    },
+    {
+      "epoch": 0.744,
+      "grad_norm": 0.2258203774690628,
+      "learning_rate": 3e-06,
+      "loss": -0.0302,
+      "policy/loss": 0.09992401748101543,
+      "step": 1395
+    },
+    {
+      "epoch": 0.7445333333333334,
+      "grad_norm": 0.3342057466506958,
+      "learning_rate": 3e-06,
+      "loss": -0.1267,
+      "policy/loss": 0.04378986099168358,
+      "step": 1396
+    },
+    {
+      "completion_length": 242.17708587646484,
+      "epoch": 0.7450666666666667,
+      "grad_norm": 0.29478558897972107,
+      "learning_rate": 3e-06,
+      "loss": -0.0924,
+      "policy/loss": 0.05646946980981937,
+      "reward": 1.3020833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.5833333432674408,
+      "step": 1397
+    },
+    {
+      "epoch": 0.7456,
+      "grad_norm": 0.23381972312927246,
+      "learning_rate": 3e-06,
+      "loss": -0.0542,
+      "policy/loss": 0.28276862342722353,
+      "step": 1398
+    },
+    {
+      "epoch": 0.7461333333333333,
+      "grad_norm": 0.2831347584724426,
+      "learning_rate": 3e-06,
+      "loss": -0.093,
+      "policy/loss": 0.049726917964889594,
+      "step": 1399
+    },
+    {
+      "epoch": 0.7466666666666667,
+      "grad_norm": 0.23648378252983093,
+      "learning_rate": 3e-06,
+      "loss": -0.056,
+      "policy/loss": 0.2800057393967279,
+      "step": 1400
+    },
+    {
+      "completion_length": 232.52083587646484,
+      "epoch": 0.7472,
+      "grad_norm": 0.2148069143295288,
+      "learning_rate": 3e-06,
+      "loss": 0.0482,
+      "policy/loss": -0.5622831281833207,
+      "reward": 1.3229166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.6666666716337204,
+      "step": 1401
+    },
+    {
+      "epoch": 0.7477333333333334,
+      "grad_norm": 0.2182644009590149,
+      "learning_rate": 3e-06,
+      "loss": -0.0256,
+      "policy/loss": -0.5405060958201595,
+      "step": 1402
+    },
+    {
+      "epoch": 0.7482666666666666,
+      "grad_norm": 0.2418496161699295,
+      "learning_rate": 3e-06,
+      "loss": 0.0474,
+      "policy/loss": -0.565476566308881,
+      "step": 1403
+    },
+    {
+      "epoch": 0.7488,
+      "grad_norm": 0.22884905338287354,
+      "learning_rate": 3e-06,
+      "loss": -0.027,
+      "policy/loss": -0.5397587132089168,
+      "step": 1404
+    },
+    {
+      "completion_length": 219.53125762939453,
+      "epoch": 0.7493333333333333,
+      "grad_norm": 0.29687803983688354,
+      "learning_rate": 3e-06,
+      "loss": -0.0546,
+      "policy/loss": -0.09200913355922324,
+      "reward": 1.6614583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.9166666567325592,
+      "step": 1405
+    },
+    {
+      "epoch": 0.7498666666666667,
+      "grad_norm": 0.23891812562942505,
+      "learning_rate": 3e-06,
+      "loss": 0.0556,
+      "policy/loss": -0.24493174502489268,
+      "step": 1406
+    },
+    {
+      "epoch": 0.7504,
+      "grad_norm": 0.32391467690467834,
+      "learning_rate": 3e-06,
+      "loss": -0.0556,
+      "policy/loss": -0.09108018768917248,
+      "step": 1407
+    },
+    {
+      "epoch": 0.7509333333333333,
+      "grad_norm": 0.3046684265136719,
+      "learning_rate": 3e-06,
+      "loss": 0.0545,
+      "policy/loss": -0.2424867178469512,
+      "step": 1408
+    },
+    {
+      "completion_length": 238.0729217529297,
+      "epoch": 0.7514666666666666,
+      "grad_norm": 0.2036328762769699,
+      "learning_rate": 3e-06,
+      "loss": 0.1096,
+      "policy/loss": -0.08470937609672546,
+      "reward": 1.6354167461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.9791666567325592,
+      "step": 1409
+    },
+    {
+      "epoch": 0.752,
+      "grad_norm": 0.22572286427021027,
+      "learning_rate": 3e-06,
+      "loss": 0.013,
+      "policy/loss": -0.21442981902509928,
+      "step": 1410
+    },
+    {
+      "epoch": 0.7525333333333334,
+      "grad_norm": 0.19957606494426727,
+      "learning_rate": 3e-06,
+      "loss": 0.1085,
+      "policy/loss": -0.0925084799528122,
+      "step": 1411
+    },
+    {
+      "epoch": 0.7530666666666667,
+      "grad_norm": 0.21582065522670746,
+      "learning_rate": 3e-06,
+      "loss": 0.0114,
+      "policy/loss": -0.2195736039429903,
+      "step": 1412
+    },
+    {
+      "completion_length": 242.46875,
+      "epoch": 0.7536,
+      "grad_norm": 0.24079398810863495,
+      "learning_rate": 3e-06,
+      "loss": -0.068,
+      "policy/loss": 0.2525722378682431,
+      "reward": 1.5104167461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.8541666567325592,
+      "step": 1413
+    },
+    {
+      "epoch": 0.7541333333333333,
+      "grad_norm": 0.30440521240234375,
+      "learning_rate": 3e-06,
+      "loss": -0.0754,
+      "policy/loss": 0.20903510080087884,
+      "step": 1414
+    },
+    {
+      "epoch": 0.7546666666666667,
+      "grad_norm": 0.23084093630313873,
+      "learning_rate": 3e-06,
+      "loss": -0.0685,
+      "policy/loss": 0.2510019236354797,
+      "step": 1415
+    },
+    {
+      "epoch": 0.7552,
+      "grad_norm": 0.24346448481082916,
+      "learning_rate": 3e-06,
+      "loss": -0.0763,
+      "policy/loss": 0.20846483335091648,
+      "step": 1416
+    },
+    {
+      "completion_length": 231.8229217529297,
+      "epoch": 0.7557333333333334,
+      "grad_norm": 0.226708322763443,
+      "learning_rate": 3e-06,
+      "loss": 0.1787,
+      "policy/loss": 0.23430974036455154,
+      "reward": 1.4166666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 1417
+    },
+    {
+      "epoch": 0.7562666666666666,
+      "grad_norm": 0.2827814519405365,
+      "learning_rate": 3e-06,
+      "loss": 0.0469,
+      "policy/loss": 0.41972099989652634,
+      "step": 1418
+    },
+    {
+      "epoch": 0.7568,
+      "grad_norm": 0.23773667216300964,
+      "learning_rate": 3e-06,
+      "loss": 0.1782,
+      "policy/loss": 0.23007764667272568,
+      "step": 1419
+    },
+    {
+      "epoch": 0.7573333333333333,
+      "grad_norm": 0.4237174987792969,
+      "learning_rate": 3e-06,
+      "loss": 0.0458,
+      "policy/loss": 0.417067501693964,
+      "step": 1420
+    },
+    {
+      "completion_length": 221.50000762939453,
+      "epoch": 0.7578666666666667,
+      "grad_norm": 0.28902968764305115,
+      "learning_rate": 3e-06,
+      "loss": 0.1803,
+      "policy/loss": -0.06422073759397762,
+      "reward": 1.7708333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 1.0208333432674408,
+      "step": 1421
+    },
+    {
+      "epoch": 0.7584,
+      "grad_norm": 0.32592618465423584,
+      "learning_rate": 3e-06,
+      "loss": 0.1891,
+      "policy/loss": -0.22315663612126357,
+      "step": 1422
+    },
+    {
+      "epoch": 0.7589333333333333,
+      "grad_norm": 0.3028984069824219,
+      "learning_rate": 3e-06,
+      "loss": 0.1789,
+      "policy/loss": -0.0635681113048534,
+      "step": 1423
+    },
+    {
+      "epoch": 0.7594666666666666,
+      "grad_norm": 0.3153185546398163,
+      "learning_rate": 3e-06,
+      "loss": 0.1892,
+      "policy/loss": -0.2217904528571779,
+      "step": 1424
+    },
+    {
+      "completion_length": 219.80209350585938,
+      "epoch": 0.76,
+      "grad_norm": 0.22657367587089539,
+      "learning_rate": 3e-06,
+      "loss": 0.035,
+      "policy/loss": -0.15684204240958444,
+      "reward": 1.4895833432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.8333333283662796,
+      "step": 1425
+    },
+    {
+      "epoch": 0.7605333333333333,
+      "grad_norm": 0.231583833694458,
+      "learning_rate": 3e-06,
+      "loss": -0.0164,
+      "policy/loss": -0.0003948966938196463,
+      "step": 1426
+    },
+    {
+      "epoch": 0.7610666666666667,
+      "grad_norm": 0.22834500670433044,
+      "learning_rate": 3e-06,
+      "loss": 0.0336,
+      "policy/loss": -0.1578072650988016,
+      "step": 1427
+    },
+    {
+      "epoch": 0.7616,
+      "grad_norm": 0.2313060462474823,
+      "learning_rate": 3e-06,
+      "loss": -0.0169,
+      "policy/loss": -0.0015905497558286186,
+      "step": 1428
+    },
+    {
+      "completion_length": 233.92708587646484,
+      "epoch": 0.7621333333333333,
+      "grad_norm": 0.2320733517408371,
+      "learning_rate": 3e-06,
+      "loss": -0.0291,
+      "policy/loss": -0.03014116467320349,
+      "reward": 1.5937500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8750000298023224,
+      "step": 1429
+    },
+    {
+      "epoch": 0.7626666666666667,
+      "grad_norm": 0.23511485755443573,
+      "learning_rate": 3e-06,
+      "loss": -0.0558,
+      "policy/loss": -0.07561698946463835,
+      "step": 1430
+    },
+    {
+      "epoch": 0.7632,
+      "grad_norm": 0.20656125247478485,
+      "learning_rate": 3e-06,
+      "loss": -0.0297,
+      "policy/loss": -0.030929173457877823,
+      "step": 1431
+    },
+    {
+      "epoch": 0.7637333333333334,
+      "grad_norm": 0.23578087985515594,
+      "learning_rate": 3e-06,
+      "loss": -0.0573,
+      "policy/loss": -0.07659054946215349,
+      "step": 1432
+    },
+    {
+      "completion_length": 222.95833587646484,
+      "epoch": 0.7642666666666666,
+      "grad_norm": 0.18832120299339294,
+      "learning_rate": 3e-06,
+      "loss": 0.004,
+      "policy/loss": 1.3466570791820232e-07,
+      "reward": 1.6666666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.9791666865348816,
+      "step": 1433
+    },
+    {
+      "epoch": 0.7648,
+      "grad_norm": 0.21489961445331573,
+      "learning_rate": 3e-06,
+      "loss": -0.0452,
+      "policy/loss": 1.632041417565233e-07,
+      "step": 1434
+    },
+    {
+      "epoch": 0.7653333333333333,
+      "grad_norm": 0.18838705122470856,
+      "learning_rate": 3e-06,
+      "loss": 0.0027,
+      "policy/loss": 1.348680669366331e-07,
+      "step": 1435
+    },
+    {
+      "epoch": 0.7658666666666667,
+      "grad_norm": 0.19523651897907257,
+      "learning_rate": 3e-06,
+      "loss": -0.0457,
+      "policy/loss": 1.630744321801103e-07,
+      "step": 1436
+    },
+    {
+      "completion_length": 232.9166717529297,
+      "epoch": 0.7664,
+      "grad_norm": 0.44764241576194763,
+      "learning_rate": 3e-06,
+      "loss": -0.0987,
+      "policy/loss": 1.9160081166091913e-07,
+      "reward": 1.5104166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 1437
+    },
+    {
+      "epoch": 0.7669333333333334,
+      "grad_norm": 0.26054519414901733,
+      "learning_rate": 3e-06,
+      "loss": -0.0161,
+      "policy/loss": 1.804404234917456e-07,
+      "step": 1438
+    },
+    {
+      "epoch": 0.7674666666666666,
+      "grad_norm": 0.24633917212486267,
+      "learning_rate": 3e-06,
+      "loss": -0.0997,
+      "policy/loss": 1.91652887338023e-07,
+      "step": 1439
+    },
+    {
+      "epoch": 0.768,
+      "grad_norm": 0.28873756527900696,
+      "learning_rate": 3e-06,
+      "loss": -0.0176,
+      "policy/loss": 1.8055262529514948e-07,
+      "step": 1440
+    },
+    {
+      "completion_length": 224.03125,
+      "epoch": 0.7685333333333333,
+      "grad_norm": 0.2300124168395996,
+      "learning_rate": 3e-06,
+      "loss": 0.0877,
+      "policy/loss": 4.662300856494994e-08,
+      "reward": 1.6093750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.9583333432674408,
+      "step": 1441
+    },
+    {
+      "epoch": 0.7690666666666667,
+      "grad_norm": 0.2788723111152649,
+      "learning_rate": 3e-06,
+      "loss": 0.0632,
+      "policy/loss": 6.81200926777592e-08,
+      "step": 1442
+    },
+    {
+      "epoch": 0.7696,
+      "grad_norm": 0.24681226909160614,
+      "learning_rate": 3e-06,
+      "loss": 0.0871,
+      "policy/loss": 4.659449270860705e-08,
+      "step": 1443
+    },
+    {
+      "epoch": 0.7701333333333333,
+      "grad_norm": 0.2888546288013458,
+      "learning_rate": 3e-06,
+      "loss": 0.0628,
+      "policy/loss": 6.808559582793805e-08,
+      "step": 1444
+    },
+    {
+      "completion_length": 211.9791717529297,
+      "epoch": 0.7706666666666667,
+      "grad_norm": 0.3134615421295166,
+      "learning_rate": 3e-06,
+      "loss": -0.0804,
+      "policy/loss": -0.5963702538394973,
+      "reward": 1.4687500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 1445
+    },
+    {
+      "epoch": 0.7712,
+      "grad_norm": 0.22664065659046173,
+      "learning_rate": 3e-06,
+      "loss": -0.0028,
+      "policy/loss": -0.3053261660403823,
+      "step": 1446
+    },
+    {
+      "epoch": 0.7717333333333334,
+      "grad_norm": 0.31294646859169006,
+      "learning_rate": 3e-06,
+      "loss": -0.0818,
+      "policy/loss": -0.5989511348947829,
+      "step": 1447
+    },
+    {
+      "epoch": 0.7722666666666667,
+      "grad_norm": 0.2469300478696823,
+      "learning_rate": 3e-06,
+      "loss": -0.0036,
+      "policy/loss": -0.3073171698346435,
+      "step": 1448
+    },
+    {
+      "completion_length": 216.11458587646484,
+      "epoch": 0.7728,
+      "grad_norm": 0.3299027383327484,
+      "learning_rate": 3e-06,
+      "loss": -0.0115,
+      "policy/loss": 7.001160184927357e-08,
+      "reward": 1.3541666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.6666666716337204,
+      "step": 1449
+    },
+    {
+      "epoch": 0.7733333333333333,
+      "grad_norm": 0.39671584963798523,
+      "learning_rate": 3e-06,
+      "loss": 0.0963,
+      "policy/loss": 6.955040987577377e-08,
+      "step": 1450
+    },
+    {
+      "epoch": 0.7738666666666667,
+      "grad_norm": 0.31349992752075195,
+      "learning_rate": 3e-06,
+      "loss": -0.0126,
+      "policy/loss": 7.003642110703368e-08,
+      "step": 1451
+    },
+    {
+      "epoch": 0.7744,
+      "grad_norm": 0.41537678241729736,
+      "learning_rate": 3e-06,
+      "loss": 0.0954,
+      "policy/loss": 6.953280973220899e-08,
+      "step": 1452
+    },
+    {
+      "completion_length": 235.0416717529297,
+      "epoch": 0.7749333333333334,
+      "grad_norm": 0.22382020950317383,
+      "learning_rate": 3e-06,
+      "loss": 0.0993,
+      "policy/loss": 0.18652054951446218,
+      "reward": 1.4635416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 1453
+    },
+    {
+      "epoch": 0.7754666666666666,
+      "grad_norm": 0.28022223711013794,
+      "learning_rate": 3e-06,
+      "loss": 0.0111,
+      "policy/loss": 0.41894491277221135,
+      "step": 1454
+    },
+    {
+      "epoch": 0.776,
+      "grad_norm": 0.23695632815361023,
+      "learning_rate": 3e-06,
+      "loss": 0.0985,
+      "policy/loss": 0.18470439607733624,
+      "step": 1455
+    },
+    {
+      "epoch": 0.7765333333333333,
+      "grad_norm": 0.4333151578903198,
+      "learning_rate": 3e-06,
+      "loss": 0.0101,
+      "policy/loss": 0.412061679043326,
+      "step": 1456
+    },
+    {
+      "completion_length": 241.3854217529297,
+      "epoch": 0.7770666666666667,
+      "grad_norm": 0.2748801112174988,
+      "learning_rate": 3e-06,
+      "loss": -0.0938,
+      "policy/loss": 6.373131355985606e-08,
+      "reward": 1.15625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6770833134651184,
+      "rewards/correctness_reward_func_math": 0.4791666716337204,
+      "step": 1457
+    },
+    {
+      "epoch": 0.7776,
+      "grad_norm": 0.2727556824684143,
+      "learning_rate": 3e-06,
+      "loss": -0.2066,
+      "policy/loss": 6.159629606372619e-08,
+      "step": 1458
+    },
+    {
+      "epoch": 0.7781333333333333,
+      "grad_norm": 0.27327367663383484,
+      "learning_rate": 3e-06,
+      "loss": -0.0953,
+      "policy/loss": 6.376578909339514e-08,
+      "step": 1459
+    },
+    {
+      "epoch": 0.7786666666666666,
+      "grad_norm": 0.2690938115119934,
+      "learning_rate": 3e-06,
+      "loss": -0.2078,
+      "policy/loss": 6.15566264627887e-08,
+      "step": 1460
+    },
+    {
+      "completion_length": 228.80208587646484,
+      "epoch": 0.7792,
+      "grad_norm": 0.3053146004676819,
+      "learning_rate": 3e-06,
+      "loss": -0.0095,
+      "policy/loss": 1.764459582886957e-07,
+      "reward": 1.5208333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.8333333134651184,
+      "step": 1461
+    },
+    {
+      "epoch": 0.7797333333333333,
+      "grad_norm": 0.2410377562046051,
+      "learning_rate": 3e-06,
+      "loss": 0.0329,
+      "policy/loss": 1.5197498726138292e-07,
+      "step": 1462
+    },
+    {
+      "epoch": 0.7802666666666667,
+      "grad_norm": 0.2982087731361389,
+      "learning_rate": 3e-06,
+      "loss": -0.0112,
+      "policy/loss": 1.765036898859762e-07,
+      "step": 1463
+    },
+    {
+      "epoch": 0.7808,
+      "grad_norm": 0.24156102538108826,
+      "learning_rate": 3e-06,
+      "loss": 0.033,
+      "policy/loss": 1.5188329172133308e-07,
+      "step": 1464
+    },
+    {
+      "completion_length": 229.17709350585938,
+      "epoch": 0.7813333333333333,
+      "grad_norm": 0.34583744406700134,
+      "learning_rate": 3e-06,
+      "loss": 0.0458,
+      "policy/loss": 1.2007845739958611e-07,
+      "reward": 1.2708333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 1465
+    },
+    {
+      "epoch": 0.7818666666666667,
+      "grad_norm": 0.33648112416267395,
+      "learning_rate": 3e-06,
+      "loss": -0.0905,
+      "policy/loss": 1.3805854237602944e-07,
+      "step": 1466
+    },
+    {
+      "epoch": 0.7824,
+      "grad_norm": 0.3616285026073456,
+      "learning_rate": 3e-06,
+      "loss": 0.0441,
+      "policy/loss": 1.2008312566536006e-07,
+      "step": 1467
+    },
+    {
+      "epoch": 0.7829333333333334,
+      "grad_norm": 0.2931138575077057,
+      "learning_rate": 3e-06,
+      "loss": -0.0913,
+      "policy/loss": 1.3813840027410151e-07,
+      "step": 1468
+    },
+    {
+      "completion_length": 233.7604217529297,
+      "epoch": 0.7834666666666666,
+      "grad_norm": 0.2235563099384308,
+      "learning_rate": 3e-06,
+      "loss": -0.2145,
+      "policy/loss": -0.6255351528525352,
+      "reward": 1.3645833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 1469
+    },
+    {
+      "epoch": 0.784,
+      "grad_norm": 0.25993502140045166,
+      "learning_rate": 3e-06,
+      "loss": 0.1238,
+      "policy/loss": 0.5409014821052551,
+      "step": 1470
+    },
+    {
+      "epoch": 0.7845333333333333,
+      "grad_norm": 0.3943372070789337,
+      "learning_rate": 3e-06,
+      "loss": -0.2143,
+      "policy/loss": -0.6272592172026634,
+      "step": 1471
+    },
+    {
+      "epoch": 0.7850666666666667,
+      "grad_norm": 0.2183302491903305,
+      "learning_rate": 3e-06,
+      "loss": 0.1236,
+      "policy/loss": 0.5406267046928406,
+      "step": 1472
+    },
+    {
+      "completion_length": 229.86459350585938,
+      "epoch": 0.7856,
+      "grad_norm": 0.21484069526195526,
+      "learning_rate": 3e-06,
+      "loss": 0.0568,
+      "policy/loss": -0.08674247562885284,
+      "reward": 1.40625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 1473
+    },
+    {
+      "epoch": 0.7861333333333334,
+      "grad_norm": 0.24848854541778564,
+      "learning_rate": 3e-06,
+      "loss": 0.1254,
+      "policy/loss": 0.3853890299797058,
+      "step": 1474
+    },
+    {
+      "epoch": 0.7866666666666666,
+      "grad_norm": 0.21176671981811523,
+      "learning_rate": 3e-06,
+      "loss": 0.0561,
+      "policy/loss": -0.08662554621696472,
+      "step": 1475
+    },
+    {
+      "epoch": 0.7872,
+      "grad_norm": 0.3038824796676636,
+      "learning_rate": 3e-06,
+      "loss": 0.1248,
+      "policy/loss": 0.38308942317962646,
+      "step": 1476
+    },
+    {
+      "completion_length": 223.34375,
+      "epoch": 0.7877333333333333,
+      "grad_norm": 0.26722949743270874,
+      "learning_rate": 3e-06,
+      "loss": 0.0246,
+      "policy/loss": 0.20370175268286062,
+      "reward": 1.375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 1477
+    },
+    {
+      "epoch": 0.7882666666666667,
+      "grad_norm": 0.24752049148082733,
+      "learning_rate": 3e-06,
+      "loss": 0.0196,
+      "policy/loss": -0.1760247008061384,
+      "step": 1478
+    },
+    {
+      "epoch": 0.7888,
+      "grad_norm": 0.30262699723243713,
+      "learning_rate": 3e-06,
+      "loss": 0.023,
+      "policy/loss": 0.19852023158384213,
+      "step": 1479
+    },
+    {
+      "epoch": 0.7893333333333333,
+      "grad_norm": 0.25891438126564026,
+      "learning_rate": 3e-06,
+      "loss": 0.0184,
+      "policy/loss": -0.17685140233407992,
+      "step": 1480
+    },
+    {
+      "completion_length": 229.27083587646484,
+      "epoch": 0.7898666666666667,
+      "grad_norm": 0.28627729415893555,
+      "learning_rate": 3e-06,
+      "loss": 0.0742,
+      "policy/loss": 0.20151189946096792,
+      "reward": 1.6666667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.9166666567325592,
+      "step": 1481
+    },
+    {
+      "epoch": 0.7904,
+      "grad_norm": 0.2916857600212097,
+      "learning_rate": 3e-06,
+      "loss": 0.0125,
+      "policy/loss": 0.18923857539397915,
+      "step": 1482
+    },
+    {
+      "epoch": 0.7909333333333334,
+      "grad_norm": 0.2867504954338074,
+      "learning_rate": 3e-06,
+      "loss": 0.073,
+      "policy/loss": 0.20023387173648644,
+      "step": 1483
+    },
+    {
+      "epoch": 0.7914666666666667,
+      "grad_norm": 0.28887590765953064,
+      "learning_rate": 3e-06,
+      "loss": 0.0097,
+      "policy/loss": 0.1865600023813201,
+      "step": 1484
+    },
+    {
+      "completion_length": 210.52084350585938,
+      "epoch": 0.792,
+      "grad_norm": 0.3630540668964386,
+      "learning_rate": 3e-06,
+      "loss": 0.025,
+      "policy/loss": 0.15446346922797716,
+      "reward": 2.0000000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 1.3125,
+      "step": 1485
+    },
+    {
+      "epoch": 0.7925333333333333,
+      "grad_norm": 0.29208904504776,
+      "learning_rate": 3e-06,
+      "loss": -0.0177,
+      "policy/loss": -0.12746280900801565,
+      "step": 1486
+    },
+    {
+      "epoch": 0.7930666666666667,
+      "grad_norm": 0.355850487947464,
+      "learning_rate": 3e-06,
+      "loss": 0.0238,
+      "policy/loss": 0.15469896721366183,
+      "step": 1487
+    },
+    {
+      "epoch": 0.7936,
+      "grad_norm": 0.294563889503479,
+      "learning_rate": 3e-06,
+      "loss": -0.0184,
+      "policy/loss": -0.12851542710553332,
+      "step": 1488
+    },
+    {
+      "completion_length": 247.58333587646484,
+      "epoch": 0.7941333333333334,
+      "grad_norm": 0.22386613488197327,
+      "learning_rate": 3e-06,
+      "loss": 0.125,
+      "policy/loss": 0.165832730311271,
+      "reward": 1.3645833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 1489
+    },
+    {
+      "epoch": 0.7946666666666666,
+      "grad_norm": 0.260802298784256,
+      "learning_rate": 3e-06,
+      "loss": 0.0542,
+      "policy/loss": -0.4365284570010086,
+      "step": 1490
+    },
+    {
+      "epoch": 0.7952,
+      "grad_norm": 0.23278287053108215,
+      "learning_rate": 3e-06,
+      "loss": 0.1245,
+      "policy/loss": 0.1662064961333538,
+      "step": 1491
+    },
+    {
+      "epoch": 0.7957333333333333,
+      "grad_norm": 0.24929578602313995,
+      "learning_rate": 3e-06,
+      "loss": 0.0527,
+      "policy/loss": -0.43558878977498594,
+      "step": 1492
+    },
+    {
+      "completion_length": 236.6979217529297,
+      "epoch": 0.7962666666666667,
+      "grad_norm": 0.19107329845428467,
+      "learning_rate": 3e-06,
+      "loss": 0.097,
+      "policy/loss": -0.31740667167894543,
+      "reward": 1.3802083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.6666666567325592,
+      "step": 1493
+    },
+    {
+      "epoch": 0.7968,
+      "grad_norm": 0.21167097985744476,
+      "learning_rate": 3e-06,
+      "loss": 0.1178,
+      "policy/loss": -0.25231226404716267,
+      "step": 1494
+    },
+    {
+      "epoch": 0.7973333333333333,
+      "grad_norm": 0.1866505742073059,
+      "learning_rate": 3e-06,
+      "loss": 0.0959,
+      "policy/loss": -0.3182815785238198,
+      "step": 1495
+    },
+    {
+      "epoch": 0.7978666666666666,
+      "grad_norm": 0.20532774925231934,
+      "learning_rate": 3e-06,
+      "loss": 0.1177,
+      "policy/loss": -0.2509089023437099,
+      "step": 1496
+    },
+    {
+      "completion_length": 218.0729217529297,
+      "epoch": 0.7984,
+      "grad_norm": 0.4841870069503784,
+      "learning_rate": 3e-06,
+      "loss": -0.1254,
+      "policy/loss": -0.2984119247392911,
+      "reward": 1.5364583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.8541666567325592,
+      "step": 1497
+    },
+    {
+      "epoch": 0.7989333333333334,
+      "grad_norm": 0.270068496465683,
+      "learning_rate": 3e-06,
+      "loss": -0.0943,
+      "policy/loss": 0.1287558725064315,
+      "step": 1498
+    },
+    {
+      "epoch": 0.7994666666666667,
+      "grad_norm": 0.5327320098876953,
+      "learning_rate": 3e-06,
+      "loss": -0.1266,
+      "policy/loss": -0.302363742472771,
+      "step": 1499
+    },
+    {
+      "epoch": 0.8,
+      "grad_norm": 0.2931385934352875,
+      "learning_rate": 3e-06,
+      "loss": -0.0958,
+      "policy/loss": 0.125920002606577,
+      "step": 1500
+    },
+    {
+      "completion_length": 237.1041717529297,
+      "epoch": 0.8005333333333333,
+      "grad_norm": 0.2675425708293915,
+      "learning_rate": 3e-06,
+      "loss": 0.082,
+      "policy/loss": 1.1888353768085835e-07,
+      "reward": 1.1250000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.4375,
+      "step": 1501
+    },
+    {
+      "epoch": 0.8010666666666667,
+      "grad_norm": 0.2287781685590744,
+      "learning_rate": 3e-06,
+      "loss": -0.1441,
+      "policy/loss": 1.632610207025209e-07,
+      "step": 1502
+    },
+    {
+      "epoch": 0.8016,
+      "grad_norm": 0.26465946435928345,
+      "learning_rate": 3e-06,
+      "loss": 0.0812,
+      "policy/loss": 1.1887146200706411e-07,
+      "step": 1503
+    },
+    {
+      "epoch": 0.8021333333333334,
+      "grad_norm": 0.23839297890663147,
+      "learning_rate": 3e-06,
+      "loss": -0.1455,
+      "policy/loss": 1.631871171525745e-07,
+      "step": 1504
+    },
+    {
+      "completion_length": 232.21875762939453,
+      "epoch": 0.8026666666666666,
+      "grad_norm": 0.22023072838783264,
+      "learning_rate": 3e-06,
+      "loss": -0.0593,
+      "policy/loss": 1.3221033512422764e-07,
+      "reward": 1.3281250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.6458333283662796,
+      "step": 1505
+    },
+    {
+      "epoch": 0.8032,
+      "grad_norm": 0.22538237273693085,
+      "learning_rate": 3e-06,
+      "loss": -0.0587,
+      "policy/loss": 1.1394755006222113e-07,
+      "step": 1506
+    },
+    {
+      "epoch": 0.8037333333333333,
+      "grad_norm": 0.2249407023191452,
+      "learning_rate": 3e-06,
+      "loss": -0.06,
+      "policy/loss": 1.3229387363367096e-07,
+      "step": 1507
+    },
+    {
+      "epoch": 0.8042666666666667,
+      "grad_norm": 0.22744615375995636,
+      "learning_rate": 3e-06,
+      "loss": -0.0603,
+      "policy/loss": 1.1396602417335089e-07,
+      "step": 1508
+    },
+    {
+      "completion_length": 221.33334350585938,
+      "epoch": 0.8048,
+      "grad_norm": 0.23470091819763184,
+      "learning_rate": 3e-06,
+      "loss": 0.1258,
+      "policy/loss": 1.140564016566259e-07,
+      "reward": 1.739583432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 1.0208333432674408,
+      "step": 1509
+    },
+    {
+      "epoch": 0.8053333333333333,
+      "grad_norm": 0.3250493109226227,
+      "learning_rate": 3e-06,
+      "loss": 0.0682,
+      "policy/loss": 1.2018086437137754e-07,
+      "step": 1510
+    },
+    {
+      "epoch": 0.8058666666666666,
+      "grad_norm": 0.2289114147424698,
+      "learning_rate": 3e-06,
+      "loss": 0.125,
+      "policy/loss": 1.1401203181549135e-07,
+      "step": 1511
+    },
+    {
+      "epoch": 0.8064,
+      "grad_norm": 0.33166754245758057,
+      "learning_rate": 3e-06,
+      "loss": 0.0677,
+      "policy/loss": 1.2033785878884373e-07,
+      "step": 1512
+    },
+    {
+      "completion_length": 225.06250762939453,
+      "epoch": 0.8069333333333333,
+      "grad_norm": 0.24887210130691528,
+      "learning_rate": 3e-06,
+      "loss": 0.0244,
+      "policy/loss": 0.057365851763336195,
+      "reward": 1.2447916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.5625,
+      "step": 1513
+    },
+    {
+      "epoch": 0.8074666666666667,
+      "grad_norm": 0.22044576704502106,
+      "learning_rate": 3e-06,
+      "loss": -0.0132,
+      "policy/loss": 0.2323655053016438,
+      "step": 1514
+    },
+    {
+      "epoch": 0.808,
+      "grad_norm": 0.2402484267950058,
+      "learning_rate": 3e-06,
+      "loss": 0.0231,
+      "policy/loss": 0.05853400220679106,
+      "step": 1515
+    },
+    {
+      "epoch": 0.8085333333333333,
+      "grad_norm": 0.2098977416753769,
+      "learning_rate": 3e-06,
+      "loss": -0.0135,
+      "policy/loss": 0.2334092122309528,
+      "step": 1516
+    },
+    {
+      "completion_length": 228.48958587646484,
+      "epoch": 0.8090666666666667,
+      "grad_norm": 0.25367286801338196,
+      "learning_rate": 3e-06,
+      "loss": -0.051,
+      "policy/loss": 0.056095222973397085,
+      "reward": 1.6145833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8958333432674408,
+      "step": 1517
+    },
+    {
+      "epoch": 0.8096,
+      "grad_norm": 0.22441093623638153,
+      "learning_rate": 3e-06,
+      "loss": -0.0717,
+      "policy/loss": -0.38606348814380453,
+      "step": 1518
+    },
+    {
+      "epoch": 0.8101333333333334,
+      "grad_norm": 0.2578965723514557,
+      "learning_rate": 3e-06,
+      "loss": -0.0515,
+      "policy/loss": 0.05441975891159956,
+      "step": 1519
+    },
+    {
+      "epoch": 0.8106666666666666,
+      "grad_norm": 0.22622892260551453,
+      "learning_rate": 3e-06,
+      "loss": -0.0728,
+      "policy/loss": -0.38612908311850447,
+      "step": 1520
+    },
+    {
+      "completion_length": 236.98959350585938,
+      "epoch": 0.8112,
+      "grad_norm": 0.23583285510540009,
+      "learning_rate": 3e-06,
+      "loss": -0.0329,
+      "policy/loss": 1.1610359251790214e-07,
+      "reward": 1.6718750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 1.0208333432674408,
+      "step": 1521
+    },
+    {
+      "epoch": 0.8117333333333333,
+      "grad_norm": 0.24264171719551086,
+      "learning_rate": 3e-06,
+      "loss": -0.0144,
+      "policy/loss": 1.3451511193807164e-07,
+      "step": 1522
+    },
+    {
+      "epoch": 0.8122666666666667,
+      "grad_norm": 0.23644031584262848,
+      "learning_rate": 3e-06,
+      "loss": -0.033,
+      "policy/loss": 1.159647382564799e-07,
+      "step": 1523
+    },
+    {
+      "epoch": 0.8128,
+      "grad_norm": 0.21599841117858887,
+      "learning_rate": 3e-06,
+      "loss": -0.0145,
+      "policy/loss": 1.345230629112848e-07,
+      "step": 1524
+    },
+    {
+      "completion_length": 239.5416717529297,
+      "epoch": 0.8133333333333334,
+      "grad_norm": 0.3936993479728699,
+      "learning_rate": 3e-06,
+      "loss": 0.0228,
+      "policy/loss": 7.926148271053535e-08,
+      "reward": 1.2135417461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.4999999850988388,
+      "step": 1525
+    },
+    {
+      "epoch": 0.8138666666666666,
+      "grad_norm": 0.24116933345794678,
+      "learning_rate": 3e-06,
+      "loss": 0.068,
+      "policy/loss": 7.29549807232388e-08,
+      "step": 1526
+    },
+    {
+      "epoch": 0.8144,
+      "grad_norm": 0.38158905506134033,
+      "learning_rate": 3e-06,
+      "loss": 0.0201,
+      "policy/loss": 7.925866185587438e-08,
+      "step": 1527
+    },
+    {
+      "epoch": 0.8149333333333333,
+      "grad_norm": 0.24936148524284363,
+      "learning_rate": 3e-06,
+      "loss": 0.0674,
+      "policy/loss": 7.281135872005962e-08,
+      "step": 1528
+    },
+    {
+      "completion_length": 243.25000762939453,
+      "epoch": 0.8154666666666667,
+      "grad_norm": 0.2582569122314453,
+      "learning_rate": 3e-06,
+      "loss": -0.1723,
+      "policy/loss": -0.12544567152754382,
+      "reward": 1.1822917461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.5000000149011612,
+      "step": 1529
+    },
+    {
+      "epoch": 0.816,
+      "grad_norm": 0.22542160749435425,
+      "learning_rate": 3e-06,
+      "loss": -0.0863,
+      "policy/loss": -0.5358191861250141,
+      "step": 1530
+    },
+    {
+      "epoch": 0.8165333333333333,
+      "grad_norm": 0.27724745869636536,
+      "learning_rate": 3e-06,
+      "loss": -0.1737,
+      "policy/loss": -0.12950176766371158,
+      "step": 1531
+    },
+    {
+      "epoch": 0.8170666666666667,
+      "grad_norm": 0.22783741354942322,
+      "learning_rate": 3e-06,
+      "loss": -0.0862,
+      "policy/loss": -0.5387794509692014,
+      "step": 1532
+    },
+    {
+      "completion_length": 236.7291717529297,
+      "epoch": 0.8176,
+      "grad_norm": 0.248599573969841,
+      "learning_rate": 3e-06,
+      "loss": -0.0451,
+      "policy/loss": 0.15870567131787539,
+      "reward": 1.5833333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.8958333134651184,
+      "step": 1533
+    },
+    {
+      "epoch": 0.8181333333333334,
+      "grad_norm": 0.30652379989624023,
+      "learning_rate": 3e-06,
+      "loss": 0.0107,
+      "policy/loss": -0.2352578341960907,
+      "step": 1534
+    },
+    {
+      "epoch": 0.8186666666666667,
+      "grad_norm": 0.23891037702560425,
+      "learning_rate": 3e-06,
+      "loss": -0.0458,
+      "policy/loss": 0.1542439954355359,
+      "step": 1535
+    },
+    {
+      "epoch": 0.8192,
+      "grad_norm": 0.2263857126235962,
+      "learning_rate": 3e-06,
+      "loss": 0.0095,
+      "policy/loss": -0.2392547931522131,
+      "step": 1536
+    },
+    {
+      "completion_length": 235.50000762939453,
+      "epoch": 0.8197333333333333,
+      "grad_norm": 0.29914095997810364,
+      "learning_rate": 3e-06,
+      "loss": -0.0103,
+      "policy/loss": 6.419481479724709e-08,
+      "reward": 1.4375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333134651184,
+      "rewards/correctness_reward_func_math": 0.7291666865348816,
+      "step": 1537
+    },
+    {
+      "epoch": 0.8202666666666667,
+      "grad_norm": 0.2604299485683441,
+      "learning_rate": 3e-06,
+      "loss": -0.0094,
+      "policy/loss": 6.149807774136207e-08,
+      "step": 1538
+    },
+    {
+      "epoch": 0.8208,
+      "grad_norm": 0.31911030411720276,
+      "learning_rate": 3e-06,
+      "loss": -0.0124,
+      "policy/loss": 6.420548714913821e-08,
+      "step": 1539
+    },
+    {
+      "epoch": 0.8213333333333334,
+      "grad_norm": 0.2683718502521515,
+      "learning_rate": 3e-06,
+      "loss": -0.0106,
+      "policy/loss": 6.152221487809584e-08,
+      "step": 1540
+    },
+    {
+      "completion_length": 231.52084350585938,
+      "epoch": 0.8218666666666666,
+      "grad_norm": 0.22992870211601257,
+      "learning_rate": 3e-06,
+      "loss": 0.0842,
+      "policy/loss": 9.09794373171735e-08,
+      "reward": 1.3958333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7083333134651184,
+      "step": 1541
+    },
+    {
+      "epoch": 0.8224,
+      "grad_norm": 0.26302292943000793,
+      "learning_rate": 3e-06,
+      "loss": 0.1081,
+      "policy/loss": 7.577262195468393e-08,
+      "step": 1542
+    },
+    {
+      "epoch": 0.8229333333333333,
+      "grad_norm": 0.22073501348495483,
+      "learning_rate": 3e-06,
+      "loss": 0.0843,
+      "policy/loss": 9.095301933825795e-08,
+      "step": 1543
+    },
+    {
+      "epoch": 0.8234666666666667,
+      "grad_norm": 0.25036829710006714,
+      "learning_rate": 3e-06,
+      "loss": 0.1079,
+      "policy/loss": 7.572486282469981e-08,
+      "step": 1544
+    },
+    {
+      "completion_length": 232.9166717529297,
+      "epoch": 0.824,
+      "grad_norm": 0.298455148935318,
+      "learning_rate": 3e-06,
+      "loss": 0.0262,
+      "policy/loss": 8.328102651944391e-08,
+      "reward": 1.0000000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.3124999925494194,
+      "step": 1545
+    },
+    {
+      "epoch": 0.8245333333333333,
+      "grad_norm": 0.25169363617897034,
+      "learning_rate": 3e-06,
+      "loss": 0.0328,
+      "policy/loss": 8.453170252664677e-08,
+      "step": 1546
+    },
+    {
+      "epoch": 0.8250666666666666,
+      "grad_norm": 0.2919028699398041,
+      "learning_rate": 3e-06,
+      "loss": 0.024,
+      "policy/loss": 8.320990829702168e-08,
+      "step": 1547
+    },
+    {
+      "epoch": 0.8256,
+      "grad_norm": 0.25530171394348145,
+      "learning_rate": 3e-06,
+      "loss": 0.0305,
+      "policy/loss": 8.44016021517291e-08,
+      "step": 1548
+    },
+    {
+      "completion_length": 234.5625,
+      "epoch": 0.8261333333333334,
+      "grad_norm": 0.20860892534255981,
+      "learning_rate": 3e-06,
+      "loss": -0.206,
+      "policy/loss": 1.0700918551265204e-07,
+      "reward": 1.2708333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.5208333283662796,
+      "step": 1549
+    },
+    {
+      "epoch": 0.8266666666666667,
+      "grad_norm": 0.4721669554710388,
+      "learning_rate": 3e-06,
+      "loss": 0.0854,
+      "policy/loss": 9.117215071796636e-08,
+      "step": 1550
+    },
+    {
+      "epoch": 0.8272,
+      "grad_norm": 0.20391155779361725,
+      "learning_rate": 3e-06,
+      "loss": -0.2058,
+      "policy/loss": 1.0693093166480594e-07,
+      "step": 1551
+    },
+    {
+      "epoch": 0.8277333333333333,
+      "grad_norm": 0.4699164628982544,
+      "learning_rate": 3e-06,
+      "loss": 0.0843,
+      "policy/loss": 9.120116395422428e-08,
+      "step": 1552
+    },
+    {
+      "completion_length": 209.68750762939453,
+      "epoch": 0.8282666666666667,
+      "grad_norm": 0.23918430507183075,
+      "learning_rate": 3e-06,
+      "loss": -0.1097,
+      "policy/loss": -0.5554738897474323,
+      "reward": 1.8593750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 1.1458333730697632,
+      "step": 1553
+    },
+    {
+      "epoch": 0.8288,
+      "grad_norm": 0.2292567789554596,
+      "learning_rate": 3e-06,
+      "loss": -0.1273,
+      "policy/loss": -0.02586267480913307,
+      "step": 1554
+    },
+    {
+      "epoch": 0.8293333333333334,
+      "grad_norm": 0.26593971252441406,
+      "learning_rate": 3e-06,
+      "loss": -0.1122,
+      "policy/loss": -0.5544582267029128,
+      "step": 1555
+    },
+    {
+      "epoch": 0.8298666666666666,
+      "grad_norm": 0.22581873834133148,
+      "learning_rate": 3e-06,
+      "loss": -0.1294,
+      "policy/loss": -0.02656719430636656,
+      "step": 1556
+    },
+    {
+      "completion_length": 226.7291717529297,
+      "epoch": 0.8304,
+      "grad_norm": 0.1842084676027298,
+      "learning_rate": 3e-06,
+      "loss": 0.0149,
+      "policy/loss": -0.9052731841802597,
+      "reward": 1.7916666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 1.0416666567325592,
+      "step": 1557
+    },
+    {
+      "epoch": 0.8309333333333333,
+      "grad_norm": 0.23652668297290802,
+      "learning_rate": 3e-06,
+      "loss": 0.0064,
+      "policy/loss": -0.39813391864299774,
+      "step": 1558
+    },
+    {
+      "epoch": 0.8314666666666667,
+      "grad_norm": 0.19451631605625153,
+      "learning_rate": 3e-06,
+      "loss": 0.014,
+      "policy/loss": -0.9042407870292664,
+      "step": 1559
+    },
+    {
+      "epoch": 0.832,
+      "grad_norm": 0.2295360565185547,
+      "learning_rate": 3e-06,
+      "loss": 0.0062,
+      "policy/loss": -0.3980279266834259,
+      "step": 1560
+    },
+    {
+      "completion_length": 233.8541717529297,
+      "epoch": 0.8325333333333333,
+      "grad_norm": 0.400777667760849,
+      "learning_rate": 3e-06,
+      "loss": -0.0528,
+      "policy/loss": -0.13696017002729377,
+      "reward": 1.1406250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.4583333432674408,
+      "step": 1561
+    },
+    {
+      "epoch": 0.8330666666666666,
+      "grad_norm": 0.21796999871730804,
+      "learning_rate": 3e-06,
+      "loss": -0.0337,
+      "policy/loss": -0.005676353343268659,
+      "step": 1562
+    },
+    {
+      "epoch": 0.8336,
+      "grad_norm": 0.35113266110420227,
+      "learning_rate": 3e-06,
+      "loss": -0.0551,
+      "policy/loss": -0.13876780008819622,
+      "step": 1563
+    },
+    {
+      "epoch": 0.8341333333333333,
+      "grad_norm": 0.22509834170341492,
+      "learning_rate": 3e-06,
+      "loss": -0.0359,
+      "policy/loss": -0.0084985382416356,
+      "step": 1564
+    },
+    {
+      "completion_length": 239.3541717529297,
+      "epoch": 0.8346666666666667,
+      "grad_norm": 0.26092857122421265,
+      "learning_rate": 3e-06,
+      "loss": -0.0245,
+      "policy/loss": 1.1736103999737679e-07,
+      "reward": 1.3854166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 1565
+    },
+    {
+      "epoch": 0.8352,
+      "grad_norm": 0.23622435331344604,
+      "learning_rate": 3e-06,
+      "loss": 0.0978,
+      "policy/loss": 1.2844958163782394e-07,
+      "step": 1566
+    },
+    {
+      "epoch": 0.8357333333333333,
+      "grad_norm": 0.23780231177806854,
+      "learning_rate": 3e-06,
+      "loss": -0.0256,
+      "policy/loss": 1.1743267691599613e-07,
+      "step": 1567
+    },
+    {
+      "epoch": 0.8362666666666667,
+      "grad_norm": 0.304636150598526,
+      "learning_rate": 3e-06,
+      "loss": 0.0976,
+      "policy/loss": 1.2844316188420635e-07,
+      "step": 1568
+    },
+    {
+      "completion_length": 238.1666717529297,
+      "epoch": 0.8368,
+      "grad_norm": 0.2643941640853882,
+      "learning_rate": 3e-06,
+      "loss": 0.0311,
+      "policy/loss": 1.5137471365278543e-07,
+      "reward": 1.3125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.6249999850988388,
+      "step": 1569
+    },
+    {
+      "epoch": 0.8373333333333334,
+      "grad_norm": 0.35125672817230225,
+      "learning_rate": 3e-06,
+      "loss": 0.081,
+      "policy/loss": 2.0353677143702953e-07,
+      "step": 1570
+    },
+    {
+      "epoch": 0.8378666666666666,
+      "grad_norm": 0.2749411165714264,
+      "learning_rate": 3e-06,
+      "loss": 0.0305,
+      "policy/loss": 1.5110845197341405e-07,
+      "step": 1571
+    },
+    {
+      "epoch": 0.8384,
+      "grad_norm": 0.41277390718460083,
+      "learning_rate": 3e-06,
+      "loss": 0.08,
+      "policy/loss": 2.0348746687659514e-07,
+      "step": 1572
+    },
+    {
+      "completion_length": 225.58333587646484,
+      "epoch": 0.8389333333333333,
+      "grad_norm": 0.2712332010269165,
+      "learning_rate": 3e-06,
+      "loss": -0.0316,
+      "policy/loss": -0.04700004365784949,
+      "reward": 1.2916666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.6666666716337204,
+      "step": 1573
+    },
+    {
+      "epoch": 0.8394666666666667,
+      "grad_norm": 0.19527189433574677,
+      "learning_rate": 3e-06,
+      "loss": -0.0787,
+      "policy/loss": 0.13031307500003564,
+      "step": 1574
+    },
+    {
+      "epoch": 0.84,
+      "grad_norm": 0.3809891641139984,
+      "learning_rate": 3e-06,
+      "loss": -0.0335,
+      "policy/loss": -0.045511257578279896,
+      "step": 1575
+    },
+    {
+      "epoch": 0.8405333333333334,
+      "grad_norm": 0.19496490061283112,
+      "learning_rate": 3e-06,
+      "loss": -0.0789,
+      "policy/loss": 0.1291478041355809,
+      "step": 1576
+    },
+    {
+      "completion_length": 236.45833587646484,
+      "epoch": 0.8410666666666666,
+      "grad_norm": 0.2935759425163269,
+      "learning_rate": 3e-06,
+      "loss": 0.0243,
+      "policy/loss": 4.3316503273160833e-08,
+      "reward": 1.5364583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.8541666567325592,
+      "step": 1577
+    },
+    {
+      "epoch": 0.8416,
+      "grad_norm": 0.23814301192760468,
+      "learning_rate": 3e-06,
+      "loss": 0.0674,
+      "policy/loss": 4.032378830842731e-08,
+      "step": 1578
+    },
+    {
+      "epoch": 0.8421333333333333,
+      "grad_norm": 0.2769794762134552,
+      "learning_rate": 3e-06,
+      "loss": 0.0226,
+      "policy/loss": 4.3192637910749454e-08,
+      "step": 1579
+    },
+    {
+      "epoch": 0.8426666666666667,
+      "grad_norm": 0.24389687180519104,
+      "learning_rate": 3e-06,
+      "loss": 0.0656,
+      "policy/loss": 4.0257816191768825e-08,
+      "step": 1580
+    },
+    {
+      "completion_length": 227.93750762939453,
+      "epoch": 0.8432,
+      "grad_norm": 0.2222883254289627,
+      "learning_rate": 3e-06,
+      "loss": -0.1413,
+      "policy/loss": 9.262828015721425e-08,
+      "reward": 1.5468750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.8958333432674408,
+      "step": 1581
+    },
+    {
+      "epoch": 0.8437333333333333,
+      "grad_norm": 0.2785870134830475,
+      "learning_rate": 3e-06,
+      "loss": -0.0647,
+      "policy/loss": 8.050800026637717e-08,
+      "step": 1582
+    },
+    {
+      "epoch": 0.8442666666666667,
+      "grad_norm": 0.23054394125938416,
+      "learning_rate": 3e-06,
+      "loss": -0.1422,
+      "policy/loss": 9.254203447994769e-08,
+      "step": 1583
+    },
+    {
+      "epoch": 0.8448,
+      "grad_norm": 0.26376280188560486,
+      "learning_rate": 3e-06,
+      "loss": -0.066,
+      "policy/loss": 8.043916110977989e-08,
+      "step": 1584
+    },
+    {
+      "completion_length": 227.5104217529297,
+      "epoch": 0.8453333333333334,
+      "grad_norm": 0.2621057629585266,
+      "learning_rate": 3e-06,
+      "loss": 0.0576,
+      "policy/loss": -0.014789601034550515,
+      "reward": 1.5989583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.9166666716337204,
+      "step": 1585
+    },
+    {
+      "epoch": 0.8458666666666667,
+      "grad_norm": 0.40899401903152466,
+      "learning_rate": 3e-06,
+      "loss": -0.088,
+      "policy/loss": -0.3172504493139101,
+      "step": 1586
+    },
+    {
+      "epoch": 0.8464,
+      "grad_norm": 0.276965469121933,
+      "learning_rate": 3e-06,
+      "loss": 0.0564,
+      "policy/loss": -0.016903749789904055,
+      "step": 1587
+    },
+    {
+      "epoch": 0.8469333333333333,
+      "grad_norm": 0.3052978515625,
+      "learning_rate": 3e-06,
+      "loss": -0.09,
+      "policy/loss": -0.3194369862443551,
+      "step": 1588
+    },
+    {
+      "completion_length": 227.4166717529297,
+      "epoch": 0.8474666666666667,
+      "grad_norm": 0.31761449575424194,
+      "learning_rate": 3e-06,
+      "loss": -0.0155,
+      "policy/loss": -0.07746179401874542,
+      "reward": 1.1354166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.4166666716337204,
+      "step": 1589
+    },
+    {
+      "epoch": 0.848,
+      "grad_norm": 0.27598077058792114,
+      "learning_rate": 3e-06,
+      "loss": 0.0391,
+      "policy/loss": 0.04636666178703308,
+      "step": 1590
+    },
+    {
+      "epoch": 0.8485333333333334,
+      "grad_norm": 0.30607640743255615,
+      "learning_rate": 3e-06,
+      "loss": -0.0185,
+      "policy/loss": -0.08471409976482391,
+      "step": 1591
+    },
+    {
+      "epoch": 0.8490666666666666,
+      "grad_norm": 0.283943772315979,
+      "learning_rate": 3e-06,
+      "loss": 0.0371,
+      "policy/loss": 0.03659334033727646,
+      "step": 1592
+    },
+    {
+      "completion_length": 234.58333587646484,
+      "epoch": 0.8496,
+      "grad_norm": 0.18956202268600464,
+      "learning_rate": 3e-06,
+      "loss": 0.0303,
+      "policy/loss": 6.277627662143459e-08,
+      "reward": 1.0468750298023224,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.3958333283662796,
+      "step": 1593
+    },
+    {
+      "epoch": 0.8501333333333333,
+      "grad_norm": 0.24116310477256775,
+      "learning_rate": 3e-06,
+      "loss": -0.0065,
+      "policy/loss": 5.2126555516451845e-08,
+      "step": 1594
+    },
+    {
+      "epoch": 0.8506666666666667,
+      "grad_norm": 0.19997534155845642,
+      "learning_rate": 3e-06,
+      "loss": 0.0305,
+      "policy/loss": 6.284501097297834e-08,
+      "step": 1595
+    },
+    {
+      "epoch": 0.8512,
+      "grad_norm": 0.34472861886024475,
+      "learning_rate": 3e-06,
+      "loss": -0.007,
+      "policy/loss": 5.207023434650182e-08,
+      "step": 1596
+    },
+    {
+      "completion_length": 236.3229217529297,
+      "epoch": 0.8517333333333333,
+      "grad_norm": 0.23698635399341583,
+      "learning_rate": 3e-06,
+      "loss": 0.0355,
+      "policy/loss": 0.1685076033859545,
+      "reward": 1.5989583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.8541666865348816,
+      "step": 1597
+    },
+    {
+      "epoch": 0.8522666666666666,
+      "grad_norm": 0.41812440752983093,
+      "learning_rate": 3e-06,
+      "loss": 0.0186,
+      "policy/loss": -0.15211466119455963,
+      "step": 1598
+    },
+    {
+      "epoch": 0.8528,
+      "grad_norm": 0.3587959110736847,
+      "learning_rate": 3e-06,
+      "loss": 0.0348,
+      "policy/loss": 0.16555951298337845,
+      "step": 1599
+    },
+    {
+      "epoch": 0.8533333333333334,
+      "grad_norm": 0.285763680934906,
+      "learning_rate": 3e-06,
+      "loss": 0.0167,
+      "policy/loss": -0.1516250836104973,
+      "step": 1600
+    },
+    {
+      "completion_length": 235.96875762939453,
+      "epoch": 0.8538666666666667,
+      "grad_norm": 0.19996802508831024,
+      "learning_rate": 3e-06,
+      "loss": 0.214,
+      "policy/loss": 0.4422186128795147,
+      "reward": 1.1614583432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6614583432674408,
+      "rewards/correctness_reward_func_math": 0.4999999850988388,
+      "step": 1601
+    },
+    {
+      "epoch": 0.8544,
+      "grad_norm": 0.2121003419160843,
+      "learning_rate": 3e-06,
+      "loss": 0.094,
+      "policy/loss": -0.25263793393969536,
+      "step": 1602
+    },
+    {
+      "epoch": 0.8549333333333333,
+      "grad_norm": 0.2104511260986328,
+      "learning_rate": 3e-06,
+      "loss": 0.2134,
+      "policy/loss": 0.442895844578743,
+      "step": 1603
+    },
+    {
+      "epoch": 0.8554666666666667,
+      "grad_norm": 0.22102974355220795,
+      "learning_rate": 3e-06,
+      "loss": 0.0932,
+      "policy/loss": -0.2518017515540123,
+      "step": 1604
+    },
+    {
+      "completion_length": 233.12500762939453,
+      "epoch": 0.856,
+      "grad_norm": 0.3026224970817566,
+      "learning_rate": 3e-06,
+      "loss": -0.0469,
+      "policy/loss": 4.241854156816771e-08,
+      "reward": 1.2916667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.6041666716337204,
+      "step": 1605
+    },
+    {
+      "epoch": 0.8565333333333334,
+      "grad_norm": 0.209719717502594,
+      "learning_rate": 3e-06,
+      "loss": -0.0714,
+      "policy/loss": 4.642719808600759e-08,
+      "step": 1606
+    },
+    {
+      "epoch": 0.8570666666666666,
+      "grad_norm": 0.3067207634449005,
+      "learning_rate": 3e-06,
+      "loss": -0.0485,
+      "policy/loss": 4.2438816905132626e-08,
+      "step": 1607
+    },
+    {
+      "epoch": 0.8576,
+      "grad_norm": 0.2149738222360611,
+      "learning_rate": 3e-06,
+      "loss": -0.0726,
+      "policy/loss": 4.646490392445912e-08,
+      "step": 1608
+    },
+    {
+      "completion_length": 235.48958587646484,
+      "epoch": 0.8581333333333333,
+      "grad_norm": 0.2152761071920395,
+      "learning_rate": 3e-06,
+      "loss": -0.0625,
+      "policy/loss": 7.88280569707922e-08,
+      "reward": 1.3593750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 1609
+    },
+    {
+      "epoch": 0.8586666666666667,
+      "grad_norm": 0.2554543614387512,
+      "learning_rate": 3e-06,
+      "loss": 0.0953,
+      "policy/loss": 9.349584928486365e-08,
+      "step": 1610
+    },
+    {
+      "epoch": 0.8592,
+      "grad_norm": 0.20192818343639374,
+      "learning_rate": 3e-06,
+      "loss": -0.0624,
+      "policy/loss": 7.874835716847883e-08,
+      "step": 1611
+    },
+    {
+      "epoch": 0.8597333333333333,
+      "grad_norm": 0.21326638758182526,
+      "learning_rate": 3e-06,
+      "loss": 0.0951,
+      "policy/loss": 9.345207629962715e-08,
+      "step": 1612
+    },
+    {
+      "completion_length": 227.28125762939453,
+      "epoch": 0.8602666666666666,
+      "grad_norm": 0.27687931060791016,
+      "learning_rate": 3e-06,
+      "loss": -0.0794,
+      "policy/loss": -0.4042365103960037,
+      "reward": 1.5000000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5625,
+      "rewards/correctness_reward_func_math": 0.9374999701976776,
+      "step": 1613
+    },
+    {
+      "epoch": 0.8608,
+      "grad_norm": 0.26956576108932495,
+      "learning_rate": 3e-06,
+      "loss": -0.0437,
+      "policy/loss": -0.6071940064430237,
+      "step": 1614
+    },
+    {
+      "epoch": 0.8613333333333333,
+      "grad_norm": 0.3064441978931427,
+      "learning_rate": 3e-06,
+      "loss": -0.0805,
+      "policy/loss": -0.41215433180332184,
+      "step": 1615
+    },
+    {
+      "epoch": 0.8618666666666667,
+      "grad_norm": 0.2651016414165497,
+      "learning_rate": 3e-06,
+      "loss": -0.046,
+      "policy/loss": -0.6146432310342789,
+      "step": 1616
+    },
+    {
+      "completion_length": 239.0104217529297,
+      "epoch": 0.8624,
+      "grad_norm": 0.24939560890197754,
+      "learning_rate": 3e-06,
+      "loss": 0.0494,
+      "policy/loss": -0.13613955710241044,
+      "reward": 1.5000000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333134651184,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 1617
+    },
+    {
+      "epoch": 0.8629333333333333,
+      "grad_norm": 0.2738957107067108,
+      "learning_rate": 3e-06,
+      "loss": 0.0764,
+      "policy/loss": 0.30642315052554636,
+      "step": 1618
+    },
+    {
+      "epoch": 0.8634666666666667,
+      "grad_norm": 0.24826259911060333,
+      "learning_rate": 3e-06,
+      "loss": 0.0473,
+      "policy/loss": -0.13965152252895763,
+      "step": 1619
+    },
+    {
+      "epoch": 0.864,
+      "grad_norm": 0.273272305727005,
+      "learning_rate": 3e-06,
+      "loss": 0.0736,
+      "policy/loss": 0.30078511732014235,
+      "step": 1620
+    },
+    {
+      "completion_length": 226.83333587646484,
+      "epoch": 0.8645333333333334,
+      "grad_norm": 0.2134665548801422,
+      "learning_rate": 3e-06,
+      "loss": -0.0996,
+      "policy/loss": 1.0243273962373678e-07,
+      "reward": 1.7916667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 1.0416666567325592,
+      "step": 1621
+    },
+    {
+      "epoch": 0.8650666666666667,
+      "grad_norm": 0.30976611375808716,
+      "learning_rate": 3e-06,
+      "loss": 0.0391,
+      "policy/loss": 9.241380993785242e-08,
+      "step": 1622
+    },
+    {
+      "epoch": 0.8656,
+      "grad_norm": 0.22124892473220825,
+      "learning_rate": 3e-06,
+      "loss": -0.1009,
+      "policy/loss": 1.023893823060007e-07,
+      "step": 1623
+    },
+    {
+      "epoch": 0.8661333333333333,
+      "grad_norm": 0.29671454429626465,
+      "learning_rate": 3e-06,
+      "loss": 0.0397,
+      "policy/loss": 9.249276899936376e-08,
+      "step": 1624
+    },
+    {
+      "completion_length": 233.28125,
+      "epoch": 0.8666666666666667,
+      "grad_norm": 0.2233886569738388,
+      "learning_rate": 3e-06,
+      "loss": -0.0113,
+      "policy/loss": 0.01856838848107145,
+      "reward": 1.34375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.59375,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 1625
+    },
+    {
+      "epoch": 0.8672,
+      "grad_norm": 0.25654199719429016,
+      "learning_rate": 3e-06,
+      "loss": -0.0553,
+      "policy/loss": -0.022483103338529986,
+      "step": 1626
+    },
+    {
+      "epoch": 0.8677333333333334,
+      "grad_norm": 0.19946566224098206,
+      "learning_rate": 3e-06,
+      "loss": -0.0117,
+      "policy/loss": 0.02018511618603469,
+      "step": 1627
+    },
+    {
+      "epoch": 0.8682666666666666,
+      "grad_norm": 0.2514427900314331,
+      "learning_rate": 3e-06,
+      "loss": -0.0574,
+      "policy/loss": -0.02410023327565014,
+      "step": 1628
+    },
+    {
+      "completion_length": 230.83333587646484,
+      "epoch": 0.8688,
+      "grad_norm": 0.21675704419612885,
+      "learning_rate": 3e-06,
+      "loss": 0.1064,
+      "policy/loss": 0.12191723100841045,
+      "reward": 1.7395833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 1.0833333432674408,
+      "step": 1629
+    },
+    {
+      "epoch": 0.8693333333333333,
+      "grad_norm": 0.2398124486207962,
+      "learning_rate": 3e-06,
+      "loss": 0.1154,
+      "policy/loss": 0.338221937417984,
+      "step": 1630
+    },
+    {
+      "epoch": 0.8698666666666667,
+      "grad_norm": 0.22994005680084229,
+      "learning_rate": 3e-06,
+      "loss": 0.1057,
+      "policy/loss": 0.12110931798815727,
+      "step": 1631
+    },
+    {
+      "epoch": 0.8704,
+      "grad_norm": 0.2439468950033188,
+      "learning_rate": 3e-06,
+      "loss": 0.1146,
+      "policy/loss": 0.3338344693183899,
+      "step": 1632
+    },
+    {
+      "completion_length": 231.4791717529297,
+      "epoch": 0.8709333333333333,
+      "grad_norm": 0.27559927105903625,
+      "learning_rate": 3e-06,
+      "loss": -0.0752,
+      "policy/loss": 7.513870770026188e-08,
+      "reward": 1.4895833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.7708333432674408,
+      "step": 1633
+    },
+    {
+      "epoch": 0.8714666666666666,
+      "grad_norm": 0.22121284902095795,
+      "learning_rate": 3e-06,
+      "loss": -0.1451,
+      "policy/loss": 7.712123917258396e-08,
+      "step": 1634
+    },
+    {
+      "epoch": 0.872,
+      "grad_norm": 0.22702136635780334,
+      "learning_rate": 3e-06,
+      "loss": -0.0747,
+      "policy/loss": 7.512550581623145e-08,
+      "step": 1635
+    },
+    {
+      "epoch": 0.8725333333333334,
+      "grad_norm": 0.2200949639081955,
+      "learning_rate": 3e-06,
+      "loss": -0.1458,
+      "policy/loss": 7.712018046390767e-08,
+      "step": 1636
+    },
+    {
+      "completion_length": 230.6666717529297,
+      "epoch": 0.8730666666666667,
+      "grad_norm": 0.2449612021446228,
+      "learning_rate": 3e-06,
+      "loss": -0.0555,
+      "policy/loss": 0.6775506734848022,
+      "reward": 1.1145833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.4583333432674408,
+      "step": 1637
+    },
+    {
+      "epoch": 0.8736,
+      "grad_norm": 0.2103649228811264,
+      "learning_rate": 3e-06,
+      "loss": 0.0444,
+      "policy/loss": 0.966956116259098,
+      "step": 1638
+    },
+    {
+      "epoch": 0.8741333333333333,
+      "grad_norm": 0.24855832755565643,
+      "learning_rate": 3e-06,
+      "loss": -0.0559,
+      "policy/loss": 0.6742304563522339,
+      "step": 1639
+    },
+    {
+      "epoch": 0.8746666666666667,
+      "grad_norm": 0.21870671212673187,
+      "learning_rate": 3e-06,
+      "loss": 0.0433,
+      "policy/loss": 0.9589559808373451,
+      "step": 1640
+    },
+    {
+      "completion_length": 221.3541717529297,
+      "epoch": 0.8752,
+      "grad_norm": 0.4364570677280426,
+      "learning_rate": 3e-06,
+      "loss": -0.0278,
+      "policy/loss": 0.16911834478378296,
+      "reward": 1.5625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 1641
+    },
+    {
+      "epoch": 0.8757333333333334,
+      "grad_norm": 0.26268449425697327,
+      "learning_rate": 3e-06,
+      "loss": 0.0394,
+      "policy/loss": -0.013728753663599491,
+      "step": 1642
+    },
+    {
+      "epoch": 0.8762666666666666,
+      "grad_norm": 0.429129958152771,
+      "learning_rate": 3e-06,
+      "loss": -0.0299,
+      "policy/loss": 0.16725854575634003,
+      "step": 1643
+    },
+    {
+      "epoch": 0.8768,
+      "grad_norm": 0.2645246684551239,
+      "learning_rate": 3e-06,
+      "loss": 0.0378,
+      "policy/loss": -0.01628766395151615,
+      "step": 1644
+    },
+    {
+      "completion_length": 238.71875762939453,
+      "epoch": 0.8773333333333333,
+      "grad_norm": 0.21671688556671143,
+      "learning_rate": 3e-06,
+      "loss": 0.1446,
+      "policy/loss": 0.5168141424655914,
+      "reward": 1.197916716337204,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.479166679084301,
+      "step": 1645
+    },
+    {
+      "epoch": 0.8778666666666667,
+      "grad_norm": 0.2590569853782654,
+      "learning_rate": 3e-06,
+      "loss": 0.0923,
+      "policy/loss": 0.6993340849876404,
+      "step": 1646
+    },
+    {
+      "epoch": 0.8784,
+      "grad_norm": 0.21756073832511902,
+      "learning_rate": 3e-06,
+      "loss": 0.1428,
+      "policy/loss": 0.5085559487342834,
+      "step": 1647
+    },
+    {
+      "epoch": 0.8789333333333333,
+      "grad_norm": 0.21313205361366272,
+      "learning_rate": 3e-06,
+      "loss": 0.0921,
+      "policy/loss": 0.6987392902374268,
+      "step": 1648
+    },
+    {
+      "completion_length": 232.09375762939453,
+      "epoch": 0.8794666666666666,
+      "grad_norm": 0.25229892134666443,
+      "learning_rate": 3e-06,
+      "loss": 0.195,
+      "policy/loss": 9.129977129873623e-08,
+      "reward": 1.3645833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 1649
+    },
+    {
+      "epoch": 0.88,
+      "grad_norm": 0.19616252183914185,
+      "learning_rate": 3e-06,
+      "loss": 0.1515,
+      "policy/loss": 9.632789499391947e-08,
+      "step": 1650
+    },
+    {
+      "epoch": 0.8805333333333333,
+      "grad_norm": 0.2645419239997864,
+      "learning_rate": 3e-06,
+      "loss": 0.1945,
+      "policy/loss": 9.127210987003309e-08,
+      "step": 1651
+    },
+    {
+      "epoch": 0.8810666666666667,
+      "grad_norm": 0.2068888545036316,
+      "learning_rate": 3e-06,
+      "loss": 0.1506,
+      "policy/loss": 9.643317255836337e-08,
+      "step": 1652
+    },
+    {
+      "completion_length": 236.68750762939453,
+      "epoch": 0.8816,
+      "grad_norm": 0.271232545375824,
+      "learning_rate": 3e-06,
+      "loss": -0.0034,
+      "policy/loss": -0.5391247086762547,
+      "reward": 1.3854166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6666666567325592,
+      "step": 1653
+    },
+    {
+      "epoch": 0.8821333333333333,
+      "grad_norm": 0.187588170170784,
+      "learning_rate": 3e-06,
+      "loss": 0.0189,
+      "policy/loss": 0.12656471876287867,
+      "step": 1654
+    },
+    {
+      "epoch": 0.8826666666666667,
+      "grad_norm": 0.26371297240257263,
+      "learning_rate": 3e-06,
+      "loss": -0.0042,
+      "policy/loss": -0.538637500294719,
+      "step": 1655
+    },
+    {
+      "epoch": 0.8832,
+      "grad_norm": 0.20860472321510315,
+      "learning_rate": 3e-06,
+      "loss": 0.0177,
+      "policy/loss": 0.12417719520531278,
+      "step": 1656
+    },
+    {
+      "completion_length": 235.8229217529297,
+      "epoch": 0.8837333333333334,
+      "grad_norm": 0.25149592757225037,
+      "learning_rate": 3e-06,
+      "loss": 0.0236,
+      "policy/loss": -0.1622752845287323,
+      "reward": 1.5260417461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 1657
+    },
+    {
+      "epoch": 0.8842666666666666,
+      "grad_norm": 0.24550385773181915,
+      "learning_rate": 3e-06,
+      "loss": -0.2256,
+      "policy/loss": -0.10232875496149063,
+      "step": 1658
+    },
+    {
+      "epoch": 0.8848,
+      "grad_norm": 0.24996653199195862,
+      "learning_rate": 3e-06,
+      "loss": 0.0229,
+      "policy/loss": -0.15954869240522385,
+      "step": 1659
+    },
+    {
+      "epoch": 0.8853333333333333,
+      "grad_norm": 0.24506981670856476,
+      "learning_rate": 3e-06,
+      "loss": -0.2271,
+      "policy/loss": -0.10520045459270477,
+      "step": 1660
+    },
+    {
+      "completion_length": 221.90625,
+      "epoch": 0.8858666666666667,
+      "grad_norm": 0.303170770406723,
+      "learning_rate": 3e-06,
+      "loss": -0.0415,
+      "policy/loss": -0.19041023403406143,
+      "reward": 1.21875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.5625,
+      "step": 1661
+    },
+    {
+      "epoch": 0.8864,
+      "grad_norm": 0.2946211099624634,
+      "learning_rate": 3e-06,
+      "loss": -0.0151,
+      "policy/loss": 0.290008008480072,
+      "step": 1662
+    },
+    {
+      "epoch": 0.8869333333333334,
+      "grad_norm": 0.22310441732406616,
+      "learning_rate": 3e-06,
+      "loss": -0.0425,
+      "policy/loss": -0.19362276047468185,
+      "step": 1663
+    },
+    {
+      "epoch": 0.8874666666666666,
+      "grad_norm": 0.22984105348587036,
+      "learning_rate": 3e-06,
+      "loss": -0.0156,
+      "policy/loss": 0.2877247892320156,
+      "step": 1664
+    },
+    {
+      "completion_length": 239.6979217529297,
+      "epoch": 0.888,
+      "grad_norm": 0.2583060562610626,
+      "learning_rate": 3e-06,
+      "loss": 0.0988,
+      "policy/loss": 7.12855179330063e-08,
+      "reward": 1.2395833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.5833333283662796,
+      "step": 1665
+    },
+    {
+      "epoch": 0.8885333333333333,
+      "grad_norm": 0.2263694405555725,
+      "learning_rate": 3e-06,
+      "loss": 0.0188,
+      "policy/loss": 7.84024418720719e-08,
+      "step": 1666
+    },
+    {
+      "epoch": 0.8890666666666667,
+      "grad_norm": 0.2565484344959259,
+      "learning_rate": 3e-06,
+      "loss": 0.0976,
+      "policy/loss": 7.133604107423253e-08,
+      "step": 1667
+    },
+    {
+      "epoch": 0.8896,
+      "grad_norm": 0.21876020729541779,
+      "learning_rate": 3e-06,
+      "loss": 0.0182,
+      "policy/loss": 7.833554960257061e-08,
+      "step": 1668
+    },
+    {
+      "completion_length": 232.5416717529297,
+      "epoch": 0.8901333333333333,
+      "grad_norm": 0.32217538356781006,
+      "learning_rate": 3e-06,
+      "loss": -0.003,
+      "policy/loss": 0.06023634501914188,
+      "reward": 1.6093750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.8958333432674408,
+      "step": 1669
+    },
+    {
+      "epoch": 0.8906666666666667,
+      "grad_norm": 0.32892295718193054,
+      "learning_rate": 3e-06,
+      "loss": -0.0639,
+      "policy/loss": -0.09961779575488094,
+      "step": 1670
+    },
+    {
+      "epoch": 0.8912,
+      "grad_norm": 0.30231478810310364,
+      "learning_rate": 3e-06,
+      "loss": -0.0056,
+      "policy/loss": 0.061102083928197715,
+      "step": 1671
+    },
+    {
+      "epoch": 0.8917333333333334,
+      "grad_norm": 0.33851832151412964,
+      "learning_rate": 3e-06,
+      "loss": -0.0655,
+      "policy/loss": -0.0991077813432355,
+      "step": 1672
+    },
+    {
+      "completion_length": 238.3229217529297,
+      "epoch": 0.8922666666666667,
+      "grad_norm": 0.2461826354265213,
+      "learning_rate": 3e-06,
+      "loss": 0.0522,
+      "policy/loss": 1.2544367322675498e-07,
+      "reward": 1.3541667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.6041666716337204,
+      "step": 1673
+    },
+    {
+      "epoch": 0.8928,
+      "grad_norm": 0.29588544368743896,
+      "learning_rate": 3e-06,
+      "loss": 0.0011,
+      "policy/loss": 1.4671028480961468e-07,
+      "step": 1674
+    },
+    {
+      "epoch": 0.8933333333333333,
+      "grad_norm": 0.24166426062583923,
+      "learning_rate": 3e-06,
+      "loss": 0.0498,
+      "policy/loss": 1.254268120476354e-07,
+      "step": 1675
+    },
+    {
+      "epoch": 0.8938666666666667,
+      "grad_norm": 0.2884717881679535,
+      "learning_rate": 3e-06,
+      "loss": -0.0016,
+      "policy/loss": 1.4677442905508542e-07,
+      "step": 1676
+    },
+    {
+      "completion_length": 243.23958587646484,
+      "epoch": 0.8944,
+      "grad_norm": 0.2998407483100891,
+      "learning_rate": 3e-06,
+      "loss": -0.1349,
+      "policy/loss": 0.43864934146404266,
+      "reward": 1.6145833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8958333134651184,
+      "step": 1677
+    },
+    {
+      "epoch": 0.8949333333333334,
+      "grad_norm": 0.27067962288856506,
+      "learning_rate": 3e-06,
+      "loss": -0.2311,
+      "policy/loss": -0.21290382742881775,
+      "step": 1678
+    },
+    {
+      "epoch": 0.8954666666666666,
+      "grad_norm": 0.2980870008468628,
+      "learning_rate": 3e-06,
+      "loss": -0.1367,
+      "policy/loss": 0.43031109869480133,
+      "step": 1679
+    },
+    {
+      "epoch": 0.896,
+      "grad_norm": 0.31338202953338623,
+      "learning_rate": 3e-06,
+      "loss": -0.2329,
+      "policy/loss": -0.21936889737844467,
+      "step": 1680
+    },
+    {
+      "completion_length": 239.06250762939453,
+      "epoch": 0.8965333333333333,
+      "grad_norm": 0.27155545353889465,
+      "learning_rate": 3e-06,
+      "loss": 0.0319,
+      "policy/loss": 7.835084403495784e-08,
+      "reward": 1.2447917461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.5625000149011612,
+      "step": 1681
+    },
+    {
+      "epoch": 0.8970666666666667,
+      "grad_norm": 0.30299854278564453,
+      "learning_rate": 3e-06,
+      "loss": 0.1046,
+      "policy/loss": 9.843432557943288e-08,
+      "step": 1682
+    },
+    {
+      "epoch": 0.8976,
+      "grad_norm": 0.27774959802627563,
+      "learning_rate": 3e-06,
+      "loss": 0.0296,
+      "policy/loss": 7.854126060635735e-08,
+      "step": 1683
+    },
+    {
+      "epoch": 0.8981333333333333,
+      "grad_norm": 0.31101447343826294,
+      "learning_rate": 3e-06,
+      "loss": 0.1031,
+      "policy/loss": 9.838803904926863e-08,
+      "step": 1684
+    },
+    {
+      "completion_length": 242.77083587646484,
+      "epoch": 0.8986666666666666,
+      "grad_norm": 0.26831117272377014,
+      "learning_rate": 3e-06,
+      "loss": -0.0162,
+      "policy/loss": 0.07907460629940033,
+      "reward": 1.8437500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 1.125,
+      "step": 1685
+    },
+    {
+      "epoch": 0.8992,
+      "grad_norm": 0.2909254729747772,
+      "learning_rate": 3e-06,
+      "loss": 0.1019,
+      "policy/loss": 0.07160815596580505,
+      "step": 1686
+    },
+    {
+      "epoch": 0.8997333333333334,
+      "grad_norm": 0.2590221166610718,
+      "learning_rate": 3e-06,
+      "loss": -0.0158,
+      "policy/loss": 0.07625772058963776,
+      "step": 1687
+    },
+    {
+      "epoch": 0.9002666666666667,
+      "grad_norm": 0.26539862155914307,
+      "learning_rate": 3e-06,
+      "loss": 0.1013,
+      "policy/loss": 0.07124494016170502,
+      "step": 1688
+    },
+    {
+      "completion_length": 240.34375762939453,
+      "epoch": 0.9008,
+      "grad_norm": 0.2864621579647064,
+      "learning_rate": 3e-06,
+      "loss": -0.0184,
+      "policy/loss": 8.903674775240233e-08,
+      "reward": 1.4322917461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 1689
+    },
+    {
+      "epoch": 0.9013333333333333,
+      "grad_norm": 0.2888999581336975,
+      "learning_rate": 3e-06,
+      "loss": -0.1981,
+      "policy/loss": 7.176566541033935e-08,
+      "step": 1690
+    },
+    {
+      "epoch": 0.9018666666666667,
+      "grad_norm": 0.2698664367198944,
+      "learning_rate": 3e-06,
+      "loss": -0.0198,
+      "policy/loss": 8.910733662048642e-08,
+      "step": 1691
+    },
+    {
+      "epoch": 0.9024,
+      "grad_norm": 0.29396581649780273,
+      "learning_rate": 3e-06,
+      "loss": -0.1997,
+      "policy/loss": 7.181958494584251e-08,
+      "step": 1692
+    },
+    {
+      "completion_length": 240.0729217529297,
+      "epoch": 0.9029333333333334,
+      "grad_norm": 0.2404072880744934,
+      "learning_rate": 3e-06,
+      "loss": -0.0445,
+      "policy/loss": 1.2862176745898068e-07,
+      "reward": 1.1041666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.4166666716337204,
+      "step": 1693
+    },
+    {
+      "epoch": 0.9034666666666666,
+      "grad_norm": 0.34246888756752014,
+      "learning_rate": 3e-06,
+      "loss": -0.0348,
+      "policy/loss": 1.1952915812685205e-07,
+      "step": 1694
+    },
+    {
+      "epoch": 0.904,
+      "grad_norm": 0.24528653919696808,
+      "learning_rate": 3e-06,
+      "loss": -0.0456,
+      "policy/loss": 1.2881035615919245e-07,
+      "step": 1695
+    },
+    {
+      "epoch": 0.9045333333333333,
+      "grad_norm": 0.33364591002464294,
+      "learning_rate": 3e-06,
+      "loss": -0.0366,
+      "policy/loss": 1.19563889455776e-07,
+      "step": 1696
+    },
+    {
+      "completion_length": 242.96875762939453,
+      "epoch": 0.9050666666666667,
+      "grad_norm": 0.2236071527004242,
+      "learning_rate": 3e-06,
+      "loss": -0.0828,
+      "policy/loss": 4.728637748030451e-08,
+      "reward": 1.2604166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6770833134651184,
+      "rewards/correctness_reward_func_math": 0.5833333134651184,
+      "step": 1697
+    },
+    {
+      "epoch": 0.9056,
+      "grad_norm": 0.21659258008003235,
+      "learning_rate": 3e-06,
+      "loss": -0.0451,
+      "policy/loss": 4.531320918488291e-08,
+      "step": 1698
+    },
+    {
+      "epoch": 0.9061333333333333,
+      "grad_norm": 0.19131837785243988,
+      "learning_rate": 3e-06,
+      "loss": -0.0829,
+      "policy/loss": 4.728325997405136e-08,
+      "step": 1699
+    },
+    {
+      "epoch": 0.9066666666666666,
+      "grad_norm": 0.21293595433235168,
+      "learning_rate": 3e-06,
+      "loss": -0.0457,
+      "policy/loss": 4.529380248641246e-08,
+      "step": 1700
+    },
+    {
+      "completion_length": 246.33333587646484,
+      "epoch": 0.9072,
+      "grad_norm": 0.20057199895381927,
+      "learning_rate": 3e-06,
+      "loss": 0.0147,
+      "policy/loss": 1.0351429935440137e-07,
+      "reward": 1.3437500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 1701
+    },
+    {
+      "epoch": 0.9077333333333333,
+      "grad_norm": 0.42632734775543213,
+      "learning_rate": 3e-06,
+      "loss": 0.0367,
+      "policy/loss": 1.4869104703052471e-07,
+      "step": 1702
+    },
+    {
+      "epoch": 0.9082666666666667,
+      "grad_norm": 0.20873452723026276,
+      "learning_rate": 3e-06,
+      "loss": 0.0154,
+      "policy/loss": 1.0371120495733521e-07,
+      "step": 1703
+    },
+    {
+      "epoch": 0.9088,
+      "grad_norm": 0.23435254395008087,
+      "learning_rate": 3e-06,
+      "loss": 0.0358,
+      "policy/loss": 1.4853091201416646e-07,
+      "step": 1704
+    },
+    {
+      "completion_length": 247.9479217529297,
+      "epoch": 0.9093333333333333,
+      "grad_norm": 0.2110428661108017,
+      "learning_rate": 3e-06,
+      "loss": -0.0749,
+      "policy/loss": 7.158075732149882e-08,
+      "reward": 1.3802083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.6666666865348816,
+      "step": 1705
+    },
+    {
+      "epoch": 0.9098666666666667,
+      "grad_norm": 0.2340516597032547,
+      "learning_rate": 3e-06,
+      "loss": -0.0799,
+      "policy/loss": 9.207023410340298e-08,
+      "step": 1706
+    },
+    {
+      "epoch": 0.9104,
+      "grad_norm": 0.2301291972398758,
+      "learning_rate": 3e-06,
+      "loss": -0.0765,
+      "policy/loss": 7.158138259910629e-08,
+      "step": 1707
+    },
+    {
+      "epoch": 0.9109333333333334,
+      "grad_norm": 0.23794247210025787,
+      "learning_rate": 3e-06,
+      "loss": -0.0809,
+      "policy/loss": 9.192743632979727e-08,
+      "step": 1708
+    },
+    {
+      "completion_length": 239.6041717529297,
+      "epoch": 0.9114666666666666,
+      "grad_norm": 0.24587412178516388,
+      "learning_rate": 3e-06,
+      "loss": -0.0674,
+      "policy/loss": 0.024173570022686164,
+      "reward": 1.5520833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8333333134651184,
+      "step": 1709
+    },
+    {
+      "epoch": 0.912,
+      "grad_norm": 0.20964397490024567,
+      "learning_rate": 3e-06,
+      "loss": 0.0041,
+      "policy/loss": 0.14682974700863838,
+      "step": 1710
+    },
+    {
+      "epoch": 0.9125333333333333,
+      "grad_norm": 0.23064188659191132,
+      "learning_rate": 3e-06,
+      "loss": -0.0691,
+      "policy/loss": 0.021642261287897924,
+      "step": 1711
+    },
+    {
+      "epoch": 0.9130666666666667,
+      "grad_norm": 0.2099466770887375,
+      "learning_rate": 3e-06,
+      "loss": 0.0036,
+      "policy/loss": 0.1497002619822183,
+      "step": 1712
+    },
+    {
+      "completion_length": 233.93750762939453,
+      "epoch": 0.9136,
+      "grad_norm": 0.2273382693529129,
+      "learning_rate": 3e-06,
+      "loss": -0.0185,
+      "policy/loss": 0.3122841920489776,
+      "reward": 1.3385416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 1713
+    },
+    {
+      "epoch": 0.9141333333333334,
+      "grad_norm": 0.2829529345035553,
+      "learning_rate": 3e-06,
+      "loss": -0.1327,
+      "policy/loss": 0.18988350936710674,
+      "step": 1714
+    },
+    {
+      "epoch": 0.9146666666666666,
+      "grad_norm": 0.2386605441570282,
+      "learning_rate": 3e-06,
+      "loss": -0.0193,
+      "policy/loss": 0.31090914267031167,
+      "step": 1715
+    },
+    {
+      "epoch": 0.9152,
+      "grad_norm": 0.28457313776016235,
+      "learning_rate": 3e-06,
+      "loss": -0.1346,
+      "policy/loss": 0.1840221078138562,
+      "step": 1716
+    },
+    {
+      "completion_length": 229.0104217529297,
+      "epoch": 0.9157333333333333,
+      "grad_norm": 0.2158655971288681,
+      "learning_rate": 3e-06,
+      "loss": -0.0107,
+      "policy/loss": -0.849545926332663,
+      "reward": 1.4010416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 1717
+    },
+    {
+      "epoch": 0.9162666666666667,
+      "grad_norm": 0.2103637307882309,
+      "learning_rate": 3e-06,
+      "loss": -0.0239,
+      "policy/loss": -0.12445602922903731,
+      "step": 1718
+    },
+    {
+      "epoch": 0.9168,
+      "grad_norm": 0.2172398865222931,
+      "learning_rate": 3e-06,
+      "loss": -0.0106,
+      "policy/loss": -0.8486518567155912,
+      "step": 1719
+    },
+    {
+      "epoch": 0.9173333333333333,
+      "grad_norm": 0.2128305733203888,
+      "learning_rate": 3e-06,
+      "loss": -0.0239,
+      "policy/loss": -0.12501823520329225,
+      "step": 1720
+    },
+    {
+      "completion_length": 232.23959350585938,
+      "epoch": 0.9178666666666667,
+      "grad_norm": 0.26258939504623413,
+      "learning_rate": 3e-06,
+      "loss": 0.0478,
+      "policy/loss": 7.588662143120928e-08,
+      "reward": 1.3125000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 1721
+    },
+    {
+      "epoch": 0.9184,
+      "grad_norm": 0.24536272883415222,
+      "learning_rate": 3e-06,
+      "loss": 0.0781,
+      "policy/loss": 8.835302045540061e-08,
+      "step": 1722
+    },
+    {
+      "epoch": 0.9189333333333334,
+      "grad_norm": 0.24711303412914276,
+      "learning_rate": 3e-06,
+      "loss": 0.0461,
+      "policy/loss": 7.59135652117493e-08,
+      "step": 1723
+    },
+    {
+      "epoch": 0.9194666666666667,
+      "grad_norm": 0.2251894623041153,
+      "learning_rate": 3e-06,
+      "loss": 0.0774,
+      "policy/loss": 8.827931807786626e-08,
+      "step": 1724
+    },
+    {
+      "completion_length": 232.50000762939453,
+      "epoch": 0.92,
+      "grad_norm": 0.27642184495925903,
+      "learning_rate": 3e-06,
+      "loss": 0.0249,
+      "policy/loss": 5.636511879458794e-08,
+      "reward": 1.6145833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.9583333134651184,
+      "step": 1725
+    },
+    {
+      "epoch": 0.9205333333333333,
+      "grad_norm": 0.1816236972808838,
+      "learning_rate": 3e-06,
+      "loss": -0.0785,
+      "policy/loss": 6.393409179850096e-08,
+      "step": 1726
+    },
+    {
+      "epoch": 0.9210666666666667,
+      "grad_norm": 0.24602611362934113,
+      "learning_rate": 3e-06,
+      "loss": 0.0237,
+      "policy/loss": 5.636701594369242e-08,
+      "step": 1727
+    },
+    {
+      "epoch": 0.9216,
+      "grad_norm": 0.18849819898605347,
+      "learning_rate": 3e-06,
+      "loss": -0.079,
+      "policy/loss": 6.381564610080659e-08,
+      "step": 1728
+    },
+    {
+      "completion_length": 232.4375,
+      "epoch": 0.9221333333333334,
+      "grad_norm": 0.19965459406375885,
+      "learning_rate": 3e-06,
+      "loss": -0.0711,
+      "policy/loss": 1.3445180258031542e-07,
+      "reward": 1.4427083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.7291666865348816,
+      "step": 1729
+    },
+    {
+      "epoch": 0.9226666666666666,
+      "grad_norm": 0.6628069281578064,
+      "learning_rate": 3e-06,
+      "loss": -0.0057,
+      "policy/loss": 1.2723491948918308e-07,
+      "step": 1730
+    },
+    {
+      "epoch": 0.9232,
+      "grad_norm": 0.21434688568115234,
+      "learning_rate": 3e-06,
+      "loss": -0.0716,
+      "policy/loss": 1.3458960523848873e-07,
+      "step": 1731
+    },
+    {
+      "epoch": 0.9237333333333333,
+      "grad_norm": 0.5686859488487244,
+      "learning_rate": 3e-06,
+      "loss": -0.0052,
+      "policy/loss": 1.2726734865964318e-07,
+      "step": 1732
+    },
+    {
+      "completion_length": 205.52083587646484,
+      "epoch": 0.9242666666666667,
+      "grad_norm": 0.2591601610183716,
+      "learning_rate": 3e-06,
+      "loss": -0.1027,
+      "policy/loss": -0.1256409715861082,
+      "reward": 1.4635416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 1733
+    },
+    {
+      "epoch": 0.9248,
+      "grad_norm": 0.24851545691490173,
+      "learning_rate": 3e-06,
+      "loss": -0.0457,
+      "policy/loss": -0.03976370394229889,
+      "step": 1734
+    },
+    {
+      "epoch": 0.9253333333333333,
+      "grad_norm": 0.30714577436447144,
+      "learning_rate": 3e-06,
+      "loss": -0.1036,
+      "policy/loss": -0.1351997908204794,
+      "step": 1735
+    },
+    {
+      "epoch": 0.9258666666666666,
+      "grad_norm": 0.2471817433834076,
+      "learning_rate": 3e-06,
+      "loss": -0.0461,
+      "policy/loss": -0.041699960827827454,
+      "step": 1736
+    },
+    {
+      "completion_length": 235.53125,
+      "epoch": 0.9264,
+      "grad_norm": 0.23034460842609406,
+      "learning_rate": 3e-06,
+      "loss": -0.0504,
+      "policy/loss": 1.1360359408740806e-07,
+      "reward": 1.1875000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.4999999850988388,
+      "step": 1737
+    },
+    {
+      "epoch": 0.9269333333333334,
+      "grad_norm": 0.303714394569397,
+      "learning_rate": 3e-06,
+      "loss": -0.0664,
+      "policy/loss": 9.250727828202798e-08,
+      "step": 1738
+    },
+    {
+      "epoch": 0.9274666666666667,
+      "grad_norm": 0.25085923075675964,
+      "learning_rate": 3e-06,
+      "loss": -0.0501,
+      "policy/loss": 1.1361890983607736e-07,
+      "step": 1739
+    },
+    {
+      "epoch": 0.928,
+      "grad_norm": 0.38232892751693726,
+      "learning_rate": 3e-06,
+      "loss": -0.0681,
+      "policy/loss": 9.253830768329863e-08,
+      "step": 1740
+    },
+    {
+      "completion_length": 225.73958587646484,
+      "epoch": 0.9285333333333333,
+      "grad_norm": 0.4197149872779846,
+      "learning_rate": 3e-06,
+      "loss": -0.02,
+      "policy/loss": -0.21172731531331834,
+      "reward": 1.4791667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7916666865348816,
+      "step": 1741
+    },
+    {
+      "epoch": 0.9290666666666667,
+      "grad_norm": 0.3365474343299866,
+      "learning_rate": 3e-06,
+      "loss": -0.1008,
+      "policy/loss": 0.23319085988508448,
+      "step": 1742
+    },
+    {
+      "epoch": 0.9296,
+      "grad_norm": 0.41989174485206604,
+      "learning_rate": 3e-06,
+      "loss": -0.0209,
+      "policy/loss": -0.21227827090324425,
+      "step": 1743
+    },
+    {
+      "epoch": 0.9301333333333334,
+      "grad_norm": 0.35699397325515747,
+      "learning_rate": 3e-06,
+      "loss": -0.1018,
+      "policy/loss": 0.23619664761505277,
+      "step": 1744
+    },
+    {
+      "completion_length": 239.8541717529297,
+      "epoch": 0.9306666666666666,
+      "grad_norm": 0.2095348984003067,
+      "learning_rate": 3e-06,
+      "loss": 0.09,
+      "policy/loss": -0.1285786665861881,
+      "reward": 1.3489583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.6666666567325592,
+      "step": 1745
+    },
+    {
+      "epoch": 0.9312,
+      "grad_norm": 0.3008284270763397,
+      "learning_rate": 3e-06,
+      "loss": 0.0129,
+      "policy/loss": 0.27860993327978534,
+      "step": 1746
+    },
+    {
+      "epoch": 0.9317333333333333,
+      "grad_norm": 0.211646169424057,
+      "learning_rate": 3e-06,
+      "loss": 0.0895,
+      "policy/loss": -0.12839228287187865,
+      "step": 1747
+    },
+    {
+      "epoch": 0.9322666666666667,
+      "grad_norm": 0.2957710027694702,
+      "learning_rate": 3e-06,
+      "loss": 0.0112,
+      "policy/loss": 0.2763008194729011,
+      "step": 1748
+    },
+    {
+      "completion_length": 227.5416717529297,
+      "epoch": 0.9328,
+      "grad_norm": 0.22542643547058105,
+      "learning_rate": 3e-06,
+      "loss": -0.1201,
+      "policy/loss": 6.848667055692204e-08,
+      "reward": 1.5729167461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.9166666567325592,
+      "step": 1749
+    },
+    {
+      "epoch": 0.9333333333333333,
+      "grad_norm": 0.22313201427459717,
+      "learning_rate": 3e-06,
+      "loss": -0.2585,
+      "policy/loss": 6.334045998812599e-08,
+      "step": 1750
+    },
+    {
+      "epoch": 0.9338666666666666,
+      "grad_norm": 1.106351375579834,
+      "learning_rate": 3e-06,
+      "loss": -0.1198,
+      "policy/loss": 6.836461530213e-08,
+      "step": 1751
+    },
+    {
+      "epoch": 0.9344,
+      "grad_norm": 0.22112686932086945,
+      "learning_rate": 3e-06,
+      "loss": -0.2595,
+      "policy/loss": 6.346103376131396e-08,
+      "step": 1752
+    },
+    {
+      "completion_length": 229.40625762939453,
+      "epoch": 0.9349333333333333,
+      "grad_norm": 0.23915891349315643,
+      "learning_rate": 3e-06,
+      "loss": 0.084,
+      "policy/loss": 0.35452549546196366,
+      "reward": 1.40625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 1753
+    },
+    {
+      "epoch": 0.9354666666666667,
+      "grad_norm": 0.2852283716201782,
+      "learning_rate": 3e-06,
+      "loss": 0.0237,
+      "policy/loss": 0.11093751690663112,
+      "step": 1754
+    },
+    {
+      "epoch": 0.936,
+      "grad_norm": 0.23482486605644226,
+      "learning_rate": 3e-06,
+      "loss": 0.0828,
+      "policy/loss": 0.3520885595616594,
+      "step": 1755
+    },
+    {
+      "epoch": 0.9365333333333333,
+      "grad_norm": 0.3288273811340332,
+      "learning_rate": 3e-06,
+      "loss": 0.0229,
+      "policy/loss": 0.1097129618948518,
+      "step": 1756
+    },
+    {
+      "completion_length": 212.71875,
+      "epoch": 0.9370666666666667,
+      "grad_norm": 0.298893541097641,
+      "learning_rate": 3e-06,
+      "loss": -0.0108,
+      "policy/loss": 1.1186495285642195e-07,
+      "reward": 1.4322916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.7499999850988388,
+      "step": 1757
+    },
+    {
+      "epoch": 0.9376,
+      "grad_norm": 0.21382372081279755,
+      "learning_rate": 3e-06,
+      "loss": -0.1232,
+      "policy/loss": 1.0066438349554119e-07,
+      "step": 1758
+    },
+    {
+      "epoch": 0.9381333333333334,
+      "grad_norm": 0.2904658317565918,
+      "learning_rate": 3e-06,
+      "loss": -0.0126,
+      "policy/loss": 1.1175092140547349e-07,
+      "step": 1759
+    },
+    {
+      "epoch": 0.9386666666666666,
+      "grad_norm": 0.22952474653720856,
+      "learning_rate": 3e-06,
+      "loss": -0.1248,
+      "policy/loss": 1.0065746103293804e-07,
+      "step": 1760
+    },
+    {
+      "completion_length": 235.43750762939453,
+      "epoch": 0.9392,
+      "grad_norm": 0.25418853759765625,
+      "learning_rate": 3e-06,
+      "loss": 0.1354,
+      "policy/loss": -0.17477597972222014,
+      "reward": 1.7187501192092896,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 1.0000000298023224,
+      "step": 1761
+    },
+    {
+      "epoch": 0.9397333333333333,
+      "grad_norm": 0.2528723180294037,
+      "learning_rate": 3e-06,
+      "loss": -0.0756,
+      "policy/loss": -0.4290699620603675,
+      "step": 1762
+    },
+    {
+      "epoch": 0.9402666666666667,
+      "grad_norm": 0.24986951053142548,
+      "learning_rate": 3e-06,
+      "loss": 0.135,
+      "policy/loss": -0.17412695948100065,
+      "step": 1763
+    },
+    {
+      "epoch": 0.9408,
+      "grad_norm": 0.2398575246334076,
+      "learning_rate": 3e-06,
+      "loss": -0.078,
+      "policy/loss": -0.43161284531080213,
+      "step": 1764
+    },
+    {
+      "completion_length": 232.15625762939453,
+      "epoch": 0.9413333333333334,
+      "grad_norm": 0.2364986091852188,
+      "learning_rate": 3e-06,
+      "loss": -0.0165,
+      "policy/loss": 0.3385841688141227,
+      "reward": 1.4010416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 1765
+    },
+    {
+      "epoch": 0.9418666666666666,
+      "grad_norm": 0.22289086878299713,
+      "learning_rate": 3e-06,
+      "loss": -0.1516,
+      "policy/loss": -0.34704963862895966,
+      "step": 1766
+    },
+    {
+      "epoch": 0.9424,
+      "grad_norm": 0.23282793164253235,
+      "learning_rate": 3e-06,
+      "loss": -0.018,
+      "policy/loss": 0.3333173831924796,
+      "step": 1767
+    },
+    {
+      "epoch": 0.9429333333333333,
+      "grad_norm": 0.2190852016210556,
+      "learning_rate": 3e-06,
+      "loss": -0.1531,
+      "policy/loss": -0.3475552797317505,
+      "step": 1768
+    },
+    {
+      "completion_length": 245.125,
+      "epoch": 0.9434666666666667,
+      "grad_norm": 0.21519573032855988,
+      "learning_rate": 3e-06,
+      "loss": 0.0105,
+      "policy/loss": 0.008144470040470253,
+      "reward": 1.0572916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6197916567325592,
+      "rewards/correctness_reward_func_math": 0.4375,
+      "step": 1769
+    },
+    {
+      "epoch": 0.944,
+      "grad_norm": 0.2047829031944275,
+      "learning_rate": 3e-06,
+      "loss": -0.0534,
+      "policy/loss": -0.4377040957291456,
+      "step": 1770
+    },
+    {
+      "epoch": 0.9445333333333333,
+      "grad_norm": 0.2812556326389313,
+      "learning_rate": 3e-06,
+      "loss": 0.0085,
+      "policy/loss": 0.00803499673381225,
+      "step": 1771
+    },
+    {
+      "epoch": 0.9450666666666667,
+      "grad_norm": 0.2127678394317627,
+      "learning_rate": 3e-06,
+      "loss": -0.0542,
+      "policy/loss": -0.4375456963279376,
+      "step": 1772
+    },
+    {
+      "completion_length": 238.48958587646484,
+      "epoch": 0.9456,
+      "grad_norm": 0.2591908574104309,
+      "learning_rate": 3e-06,
+      "loss": -0.0074,
+      "policy/loss": 0.003083105160477828,
+      "reward": 1.1666667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.5416666865348816,
+      "step": 1773
+    },
+    {
+      "epoch": 0.9461333333333334,
+      "grad_norm": 0.2732800245285034,
+      "learning_rate": 3e-06,
+      "loss": 0.0279,
+      "policy/loss": -0.4974988395634483,
+      "step": 1774
+    },
+    {
+      "epoch": 0.9466666666666667,
+      "grad_norm": 0.25695475935935974,
+      "learning_rate": 3e-06,
+      "loss": -0.0097,
+      "policy/loss": 0.0023230267663372217,
+      "step": 1775
+    },
+    {
+      "epoch": 0.9472,
+      "grad_norm": 0.28710684180259705,
+      "learning_rate": 3e-06,
+      "loss": 0.0273,
+      "policy/loss": -0.49989849278833987,
+      "step": 1776
+    },
+    {
+      "completion_length": 225.52084350585938,
+      "epoch": 0.9477333333333333,
+      "grad_norm": 0.3070966303348541,
+      "learning_rate": 3e-06,
+      "loss": -0.0428,
+      "policy/loss": 7.384606171001451e-08,
+      "reward": 1.375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 1777
+    },
+    {
+      "epoch": 0.9482666666666667,
+      "grad_norm": 0.26500314474105835,
+      "learning_rate": 3e-06,
+      "loss": 0.0561,
+      "policy/loss": 7.682925939889174e-08,
+      "step": 1778
+    },
+    {
+      "epoch": 0.9488,
+      "grad_norm": 0.30891451239585876,
+      "learning_rate": 3e-06,
+      "loss": -0.0438,
+      "policy/loss": 7.389900069654232e-08,
+      "step": 1779
+    },
+    {
+      "epoch": 0.9493333333333334,
+      "grad_norm": 0.26080599427223206,
+      "learning_rate": 3e-06,
+      "loss": 0.0553,
+      "policy/loss": 7.688509029435409e-08,
+      "step": 1780
+    },
+    {
+      "completion_length": 226.52083587646484,
+      "epoch": 0.9498666666666666,
+      "grad_norm": 0.23890477418899536,
+      "learning_rate": 3e-06,
+      "loss": 0.0989,
+      "policy/loss": 0.31344373524188995,
+      "reward": 1.7552083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 1.0416666567325592,
+      "step": 1781
+    },
+    {
+      "epoch": 0.9504,
+      "grad_norm": 0.2694341540336609,
+      "learning_rate": 3e-06,
+      "loss": -0.0523,
+      "policy/loss": 0.5277764201164246,
+      "step": 1782
+    },
+    {
+      "epoch": 0.9509333333333333,
+      "grad_norm": 0.22837823629379272,
+      "learning_rate": 3e-06,
+      "loss": 0.097,
+      "policy/loss": 0.31098122149705887,
+      "step": 1783
+    },
+    {
+      "epoch": 0.9514666666666667,
+      "grad_norm": 0.25884103775024414,
+      "learning_rate": 3e-06,
+      "loss": -0.0529,
+      "policy/loss": 0.5296276062726974,
+      "step": 1784
+    },
+    {
+      "completion_length": 222.4166717529297,
+      "epoch": 0.952,
+      "grad_norm": 0.28113067150115967,
+      "learning_rate": 3e-06,
+      "loss": -0.0758,
+      "policy/loss": 0.13317452425655674,
+      "reward": 1.4166666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7291666716337204,
+      "step": 1785
+    },
+    {
+      "epoch": 0.9525333333333333,
+      "grad_norm": 0.37620124220848083,
+      "learning_rate": 3e-06,
+      "loss": 0.1186,
+      "policy/loss": 0.3432514306850969,
+      "step": 1786
+    },
+    {
+      "epoch": 0.9530666666666666,
+      "grad_norm": 0.3358129560947418,
+      "learning_rate": 3e-06,
+      "loss": -0.0773,
+      "policy/loss": 0.12842916005195804,
+      "step": 1787
+    },
+    {
+      "epoch": 0.9536,
+      "grad_norm": 0.2949890196323395,
+      "learning_rate": 3e-06,
+      "loss": 0.1181,
+      "policy/loss": 0.34143909203526945,
+      "step": 1788
+    },
+    {
+      "completion_length": 232.58333587646484,
+      "epoch": 0.9541333333333334,
+      "grad_norm": 0.318055659532547,
+      "learning_rate": 3e-06,
+      "loss": 0.0925,
+      "policy/loss": -0.3146296183257604,
+      "reward": 1.6614583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.9166666865348816,
+      "step": 1789
+    },
+    {
+      "epoch": 0.9546666666666667,
+      "grad_norm": 0.3287056088447571,
+      "learning_rate": 3e-06,
+      "loss": -0.0726,
+      "policy/loss": 0.06790008153582505,
+      "step": 1790
+    },
+    {
+      "epoch": 0.9552,
+      "grad_norm": 0.36514991521835327,
+      "learning_rate": 3e-06,
+      "loss": 0.0924,
+      "policy/loss": -0.31527123248951483,
+      "step": 1791
+    },
+    {
+      "epoch": 0.9557333333333333,
+      "grad_norm": 0.3766613304615021,
+      "learning_rate": 3e-06,
+      "loss": -0.0733,
+      "policy/loss": 0.06957311693850343,
+      "step": 1792
+    },
+    {
+      "completion_length": 234.1354217529297,
+      "epoch": 0.9562666666666667,
+      "grad_norm": 0.2321784496307373,
+      "learning_rate": 3e-06,
+      "loss": 0.0236,
+      "policy/loss": -0.3886779323220253,
+      "reward": 1.5625000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.8124999850988388,
+      "step": 1793
+    },
+    {
+      "epoch": 0.9568,
+      "grad_norm": 0.2039032131433487,
+      "learning_rate": 3e-06,
+      "loss": -0.1772,
+      "policy/loss": -0.5317534506320953,
+      "step": 1794
+    },
+    {
+      "epoch": 0.9573333333333334,
+      "grad_norm": 0.23955093324184418,
+      "learning_rate": 3e-06,
+      "loss": 0.0233,
+      "policy/loss": -0.3886084631085396,
+      "step": 1795
+    },
+    {
+      "epoch": 0.9578666666666666,
+      "grad_norm": 0.2120412439107895,
+      "learning_rate": 3e-06,
+      "loss": -0.1789,
+      "policy/loss": -0.5389872342348099,
+      "step": 1796
+    },
+    {
+      "completion_length": 232.8541717529297,
+      "epoch": 0.9584,
+      "grad_norm": 0.2811206579208374,
+      "learning_rate": 3e-06,
+      "loss": 0.1478,
+      "policy/loss": 0.05782452264454463,
+      "reward": 1.4895833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.770833358168602,
+      "step": 1797
+    },
+    {
+      "epoch": 0.9589333333333333,
+      "grad_norm": 0.24529050290584564,
+      "learning_rate": 3e-06,
+      "loss": 0.1608,
+      "policy/loss": 0.1294601951225438,
+      "step": 1798
+    },
+    {
+      "epoch": 0.9594666666666667,
+      "grad_norm": 0.24067190289497375,
+      "learning_rate": 3e-06,
+      "loss": 0.1477,
+      "policy/loss": 0.05791478636619729,
+      "step": 1799
+    },
+    {
+      "epoch": 0.96,
+      "grad_norm": 0.23414868116378784,
+      "learning_rate": 3e-06,
+      "loss": 0.1585,
+      "policy/loss": 0.1284951066380664,
+      "step": 1800
+    },
+    {
+      "completion_length": 231.5729217529297,
+      "epoch": 0.9605333333333334,
+      "grad_norm": 0.2620799243450165,
+      "learning_rate": 3e-06,
+      "loss": -0.0661,
+      "policy/loss": 0.6954539800536423,
+      "reward": 1.0468750298023224,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.3958333283662796,
+      "step": 1801
+    },
+    {
+      "epoch": 0.9610666666666666,
+      "grad_norm": 0.4195791184902191,
+      "learning_rate": 3e-06,
+      "loss": -0.0241,
+      "policy/loss": 0.4831081987508412,
+      "step": 1802
+    },
+    {
+      "epoch": 0.9616,
+      "grad_norm": 0.25415971875190735,
+      "learning_rate": 3e-06,
+      "loss": -0.0667,
+      "policy/loss": 0.6937057756956762,
+      "step": 1803
+    },
+    {
+      "epoch": 0.9621333333333333,
+      "grad_norm": 0.41309624910354614,
+      "learning_rate": 3e-06,
+      "loss": -0.0253,
+      "policy/loss": 0.48578319545038795,
+      "step": 1804
+    },
+    {
+      "completion_length": 223.17708587646484,
+      "epoch": 0.9626666666666667,
+      "grad_norm": 0.20356211066246033,
+      "learning_rate": 3e-06,
+      "loss": 0.0603,
+      "policy/loss": -0.08036542683839798,
+      "reward": 1.6614583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.9791666865348816,
+      "step": 1805
+    },
+    {
+      "epoch": 0.9632,
+      "grad_norm": 0.1729261875152588,
+      "learning_rate": 3e-06,
+      "loss": -0.021,
+      "policy/loss": -0.1654629111289978,
+      "step": 1806
+    },
+    {
+      "epoch": 0.9637333333333333,
+      "grad_norm": 0.22402849793434143,
+      "learning_rate": 3e-06,
+      "loss": 0.0599,
+      "policy/loss": -0.08013971149921417,
+      "step": 1807
+    },
+    {
+      "epoch": 0.9642666666666667,
+      "grad_norm": 0.17426282167434692,
+      "learning_rate": 3e-06,
+      "loss": -0.0218,
+      "policy/loss": -0.16743619740009308,
+      "step": 1808
+    },
+    {
+      "completion_length": 225.86458587646484,
+      "epoch": 0.9648,
+      "grad_norm": 0.45023536682128906,
+      "learning_rate": 3e-06,
+      "loss": -0.1097,
+      "policy/loss": 0.17380893834333477,
+      "reward": 1.7708333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 1.0208333134651184,
+      "step": 1809
+    },
+    {
+      "epoch": 0.9653333333333334,
+      "grad_norm": 0.4109441637992859,
+      "learning_rate": 3e-06,
+      "loss": 0.0545,
+      "policy/loss": 0.0184552973191785,
+      "step": 1810
+    },
+    {
+      "epoch": 0.9658666666666667,
+      "grad_norm": 0.46816039085388184,
+      "learning_rate": 3e-06,
+      "loss": -0.1101,
+      "policy/loss": 0.17364551719603583,
+      "step": 1811
+    },
+    {
+      "epoch": 0.9664,
+      "grad_norm": 0.4854927659034729,
+      "learning_rate": 3e-06,
+      "loss": 0.0507,
+      "policy/loss": 0.013740010237675904,
+      "step": 1812
+    },
+    {
+      "completion_length": 214.18750762939453,
+      "epoch": 0.9669333333333333,
+      "grad_norm": 0.47048643231391907,
+      "learning_rate": 3e-06,
+      "loss": -0.071,
+      "policy/loss": 0.3029606118798256,
+      "reward": 1.5416666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.8541666567325592,
+      "step": 1813
+    },
+    {
+      "epoch": 0.9674666666666667,
+      "grad_norm": 0.33533647656440735,
+      "learning_rate": 3e-06,
+      "loss": -0.0108,
+      "policy/loss": -0.05269334465265274,
+      "step": 1814
+    },
+    {
+      "epoch": 0.968,
+      "grad_norm": 0.5453078746795654,
+      "learning_rate": 3e-06,
+      "loss": -0.0748,
+      "policy/loss": 0.2931874915957451,
+      "step": 1815
+    },
+    {
+      "epoch": 0.9685333333333334,
+      "grad_norm": 0.3359359800815582,
+      "learning_rate": 3e-06,
+      "loss": -0.0114,
+      "policy/loss": -0.05105777084827423,
+      "step": 1816
+    },
+    {
+      "completion_length": 222.0625,
+      "epoch": 0.9690666666666666,
+      "grad_norm": 0.36759936809539795,
+      "learning_rate": 3e-06,
+      "loss": 0.0213,
+      "policy/loss": 1.0624130553082978e-07,
+      "reward": 1.2708333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.5833333283662796,
+      "step": 1817
+    },
+    {
+      "epoch": 0.9696,
+      "grad_norm": 0.33086007833480835,
+      "learning_rate": 3e-06,
+      "loss": 0.1336,
+      "policy/loss": 1.0152781371175479e-07,
+      "step": 1818
+    },
+    {
+      "epoch": 0.9701333333333333,
+      "grad_norm": 0.3851078152656555,
+      "learning_rate": 3e-06,
+      "loss": 0.0231,
+      "policy/loss": 1.0614347800697033e-07,
+      "step": 1819
+    },
+    {
+      "epoch": 0.9706666666666667,
+      "grad_norm": 0.31711897253990173,
+      "learning_rate": 3e-06,
+      "loss": 0.1329,
+      "policy/loss": 1.0147982010266787e-07,
+      "step": 1820
+    },
+    {
+      "completion_length": 235.9375,
+      "epoch": 0.9712,
+      "grad_norm": 0.23699147999286652,
+      "learning_rate": 3e-06,
+      "loss": -0.0429,
+      "policy/loss": 7.545333247094277e-08,
+      "reward": 0.8437500298023224,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.1874999962747097,
+      "step": 1821
+    },
+    {
+      "epoch": 0.9717333333333333,
+      "grad_norm": 0.3791571855545044,
+      "learning_rate": 3e-06,
+      "loss": 0.1009,
+      "policy/loss": 7.297540705053507e-08,
+      "step": 1822
+    },
+    {
+      "epoch": 0.9722666666666666,
+      "grad_norm": 0.23271100223064423,
+      "learning_rate": 3e-06,
+      "loss": -0.0442,
+      "policy/loss": 7.552688430223498e-08,
+      "step": 1823
+    },
+    {
+      "epoch": 0.9728,
+      "grad_norm": 0.37420129776000977,
+      "learning_rate": 3e-06,
+      "loss": 0.0999,
+      "policy/loss": 7.291838599599032e-08,
+      "step": 1824
+    },
+    {
+      "completion_length": 240.86459350585938,
+      "epoch": 0.9733333333333334,
+      "grad_norm": 0.264046311378479,
+      "learning_rate": 3e-06,
+      "loss": -0.1849,
+      "policy/loss": 2.383563035834868e-08,
+      "reward": 1.1875000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.4999999850988388,
+      "step": 1825
+    },
+    {
+      "epoch": 0.9738666666666667,
+      "grad_norm": 0.2967522144317627,
+      "learning_rate": 3e-06,
+      "loss": -0.0095,
+      "policy/loss": 2.7948525271881408e-08,
+      "step": 1826
+    },
+    {
+      "epoch": 0.9744,
+      "grad_norm": 0.2949065566062927,
+      "learning_rate": 3e-06,
+      "loss": -0.1859,
+      "policy/loss": 2.381131558593097e-08,
+      "step": 1827
+    },
+    {
+      "epoch": 0.9749333333333333,
+      "grad_norm": 0.31684401631355286,
+      "learning_rate": 3e-06,
+      "loss": -0.0093,
+      "policy/loss": 2.7970827432000078e-08,
+      "step": 1828
+    },
+    {
+      "completion_length": 220.5104217529297,
+      "epoch": 0.9754666666666667,
+      "grad_norm": 0.24179211258888245,
+      "learning_rate": 3e-06,
+      "loss": -0.1153,
+      "policy/loss": -0.14979112897970026,
+      "reward": 1.3125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 1829
+    },
+    {
+      "epoch": 0.976,
+      "grad_norm": 0.33733657002449036,
+      "learning_rate": 3e-06,
+      "loss": 0.0812,
+      "policy/loss": 0.2635783915737875,
+      "step": 1830
+    },
+    {
+      "epoch": 0.9765333333333334,
+      "grad_norm": 0.22971723973751068,
+      "learning_rate": 3e-06,
+      "loss": -0.1167,
+      "policy/loss": -0.15128252660851516,
+      "step": 1831
+    },
+    {
+      "epoch": 0.9770666666666666,
+      "grad_norm": 0.25276127457618713,
+      "learning_rate": 3e-06,
+      "loss": 0.0809,
+      "policy/loss": 0.2600558166936686,
+      "step": 1832
+    },
+    {
+      "completion_length": 225.4166717529297,
+      "epoch": 0.9776,
+      "grad_norm": 0.29505252838134766,
+      "learning_rate": 3e-06,
+      "loss": 0.1256,
+      "policy/loss": 0.5153273642063141,
+      "reward": 1.6145833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8958333432674408,
+      "step": 1833
+    },
+    {
+      "epoch": 0.9781333333333333,
+      "grad_norm": 0.23646466434001923,
+      "learning_rate": 3e-06,
+      "loss": 0.1123,
+      "policy/loss": 0.9205298200249672,
+      "step": 1834
+    },
+    {
+      "epoch": 0.9786666666666667,
+      "grad_norm": 0.29883521795272827,
+      "learning_rate": 3e-06,
+      "loss": 0.1244,
+      "policy/loss": 0.5110231339931488,
+      "step": 1835
+    },
+    {
+      "epoch": 0.9792,
+      "grad_norm": 0.2482961118221283,
+      "learning_rate": 3e-06,
+      "loss": 0.1122,
+      "policy/loss": 0.9230922237038612,
+      "step": 1836
+    },
+    {
+      "completion_length": 239.8229217529297,
+      "epoch": 0.9797333333333333,
+      "grad_norm": 0.4063987731933594,
+      "learning_rate": 3e-06,
+      "loss": 0.0545,
+      "policy/loss": -0.2943841889500618,
+      "reward": 1.3229166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6041666567325592,
+      "step": 1837
+    },
+    {
+      "epoch": 0.9802666666666666,
+      "grad_norm": 0.3355572521686554,
+      "learning_rate": 3e-06,
+      "loss": 0.065,
+      "policy/loss": 0.2476368322968483,
+      "step": 1838
+    },
+    {
+      "epoch": 0.9808,
+      "grad_norm": 0.38188087940216064,
+      "learning_rate": 3e-06,
+      "loss": 0.0545,
+      "policy/loss": -0.29703371226787567,
+      "step": 1839
+    },
+    {
+      "epoch": 0.9813333333333333,
+      "grad_norm": 0.31338027119636536,
+      "learning_rate": 3e-06,
+      "loss": 0.0626,
+      "policy/loss": 0.239364355802536,
+      "step": 1840
+    },
+    {
+      "completion_length": 227.1666717529297,
+      "epoch": 0.9818666666666667,
+      "grad_norm": 0.24532976746559143,
+      "learning_rate": 3e-06,
+      "loss": 0.0212,
+      "policy/loss": 0.03318903394137607,
+      "reward": 1.369791716337204,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.6875000149011612,
+      "step": 1841
+    },
+    {
+      "epoch": 0.9824,
+      "grad_norm": 0.26667022705078125,
+      "learning_rate": 3e-06,
+      "loss": -0.0047,
+      "policy/loss": 0.3021068530856965,
+      "step": 1842
+    },
+    {
+      "epoch": 0.9829333333333333,
+      "grad_norm": 0.21796970069408417,
+      "learning_rate": 3e-06,
+      "loss": 0.0207,
+      "policy/loss": 0.032773999330871106,
+      "step": 1843
+    },
+    {
+      "epoch": 0.9834666666666667,
+      "grad_norm": 0.2729778289794922,
+      "learning_rate": 3e-06,
+      "loss": -0.0061,
+      "policy/loss": 0.3035094394691207,
+      "step": 1844
+    },
+    {
+      "completion_length": 223.1354217529297,
+      "epoch": 0.984,
+      "grad_norm": 0.4042705297470093,
+      "learning_rate": 3e-06,
+      "loss": 0.0024,
+      "policy/loss": -0.30052437399808696,
+      "reward": 1.7291666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333134651184,
+      "rewards/correctness_reward_func_math": 1.0208333432674408,
+      "step": 1845
+    },
+    {
+      "epoch": 0.9845333333333334,
+      "grad_norm": 0.3013225495815277,
+      "learning_rate": 3e-06,
+      "loss": -0.0404,
+      "policy/loss": -0.29405705052917597,
+      "step": 1846
+    },
+    {
+      "epoch": 0.9850666666666666,
+      "grad_norm": 0.40077468752861023,
+      "learning_rate": 3e-06,
+      "loss": -0.0005,
+      "policy/loss": -0.303297688774812,
+      "step": 1847
+    },
+    {
+      "epoch": 0.9856,
+      "grad_norm": 0.30789828300476074,
+      "learning_rate": 3e-06,
+      "loss": -0.043,
+      "policy/loss": -0.2981266171410937,
+      "step": 1848
+    },
+    {
+      "completion_length": 227.90625762939453,
+      "epoch": 0.9861333333333333,
+      "grad_norm": 0.2361096888780594,
+      "learning_rate": 3e-06,
+      "loss": -0.041,
+      "policy/loss": -0.11405044794082642,
+      "reward": 1.25,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 1849
+    },
+    {
+      "epoch": 0.9866666666666667,
+      "grad_norm": 0.1850796490907669,
+      "learning_rate": 3e-06,
+      "loss": 0.048,
+      "policy/loss": -0.186761736869812,
+      "step": 1850
+    },
+    {
+      "epoch": 0.9872,
+      "grad_norm": 0.23259514570236206,
+      "learning_rate": 3e-06,
+      "loss": -0.0414,
+      "policy/loss": -0.11812619119882584,
+      "step": 1851
+    },
+    {
+      "epoch": 0.9877333333333334,
+      "grad_norm": 0.18500453233718872,
+      "learning_rate": 3e-06,
+      "loss": 0.0469,
+      "policy/loss": -0.1871805042028427,
+      "step": 1852
+    },
+    {
+      "completion_length": 213.6041717529297,
+      "epoch": 0.9882666666666666,
+      "grad_norm": 0.3270284831523895,
+      "learning_rate": 3e-06,
+      "loss": -0.1398,
+      "policy/loss": -1.1731662154197693,
+      "reward": 1.5416666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.8541666567325592,
+      "step": 1853
+    },
+    {
+      "epoch": 0.9888,
+      "grad_norm": 0.5599705576896667,
+      "learning_rate": 3e-06,
+      "loss": -0.1015,
+      "policy/loss": -0.7747231423854828,
+      "step": 1854
+    },
+    {
+      "epoch": 0.9893333333333333,
+      "grad_norm": 0.3241273760795593,
+      "learning_rate": 3e-06,
+      "loss": -0.14,
+      "policy/loss": -1.1745698899030685,
+      "step": 1855
+    },
+    {
+      "epoch": 0.9898666666666667,
+      "grad_norm": 0.3338998258113861,
+      "learning_rate": 3e-06,
+      "loss": -0.1025,
+      "policy/loss": -0.7713559614494443,
+      "step": 1856
+    },
+    {
+      "completion_length": 223.73959350585938,
+      "epoch": 0.9904,
+      "grad_norm": 0.23343002796173096,
+      "learning_rate": 3e-06,
+      "loss": 0.0524,
+      "policy/loss": 1.0416500953511196,
+      "reward": 1.53125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 1857
+    },
+    {
+      "epoch": 0.9909333333333333,
+      "grad_norm": 0.21826617419719696,
+      "learning_rate": 3e-06,
+      "loss": 0.0322,
+      "policy/loss": 0.2477991324623332,
+      "step": 1858
+    },
+    {
+      "epoch": 0.9914666666666667,
+      "grad_norm": 0.21602608263492584,
+      "learning_rate": 3e-06,
+      "loss": 0.0515,
+      "policy/loss": 1.041455903364536,
+      "step": 1859
+    },
+    {
+      "epoch": 0.992,
+      "grad_norm": 0.21637728810310364,
+      "learning_rate": 3e-06,
+      "loss": 0.0313,
+      "policy/loss": 0.24323676950586304,
+      "step": 1860
+    },
+    {
+      "completion_length": 231.0104217529297,
+      "epoch": 0.9925333333333334,
+      "grad_norm": 0.27991583943367004,
+      "learning_rate": 3e-06,
+      "loss": -0.1027,
+      "policy/loss": 3.196645081970928e-08,
+      "reward": 1.3020833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 1861
+    },
+    {
+      "epoch": 0.9930666666666667,
+      "grad_norm": 0.29087886214256287,
+      "learning_rate": 3e-06,
+      "loss": 0.187,
+      "policy/loss": 3.898033185123495e-08,
+      "step": 1862
+    },
+    {
+      "epoch": 0.9936,
+      "grad_norm": 0.269622266292572,
+      "learning_rate": 3e-06,
+      "loss": -0.1039,
+      "policy/loss": 3.194939068862368e-08,
+      "step": 1863
+    },
+    {
+      "epoch": 0.9941333333333333,
+      "grad_norm": 0.28416091203689575,
+      "learning_rate": 3e-06,
+      "loss": 0.1883,
+      "policy/loss": 3.889959998559789e-08,
+      "step": 1864
+    },
+    {
+      "completion_length": 227.125,
+      "epoch": 0.9946666666666667,
+      "grad_norm": 0.2114567756652832,
+      "learning_rate": 3e-06,
+      "loss": -0.0715,
+      "policy/loss": 0.09602871583324202,
+      "reward": 1.4427083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 1865
+    },
+    {
+      "epoch": 0.9952,
+      "grad_norm": 0.22727060317993164,
+      "learning_rate": 3e-06,
+      "loss": -0.0641,
+      "policy/loss": 0.03508235049219621,
+      "step": 1866
+    },
+    {
+      "epoch": 0.9957333333333334,
+      "grad_norm": 0.22845426201820374,
+      "learning_rate": 3e-06,
+      "loss": -0.0728,
+      "policy/loss": 0.09589131216988989,
+      "step": 1867
+    },
+    {
+      "epoch": 0.9962666666666666,
+      "grad_norm": 0.19475124776363373,
+      "learning_rate": 3e-06,
+      "loss": -0.0644,
+      "policy/loss": 0.0343300728346243,
+      "step": 1868
+    },
+    {
+      "completion_length": 226.46875762939453,
+      "epoch": 0.9968,
+      "grad_norm": 0.26459091901779175,
+      "learning_rate": 3e-06,
+      "loss": 0.0277,
+      "policy/loss": 0.4299321807920933,
+      "reward": 1.2239583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.5416666567325592,
+      "step": 1869
+    },
+    {
+      "epoch": 0.9973333333333333,
+      "grad_norm": 0.32212573289871216,
+      "learning_rate": 3e-06,
+      "loss": 0.0784,
+      "policy/loss": 0.15804274380207062,
+      "step": 1870
+    },
+    {
+      "epoch": 0.9978666666666667,
+      "grad_norm": 0.2707771956920624,
+      "learning_rate": 3e-06,
+      "loss": 0.0265,
+      "policy/loss": 0.4280464109033346,
+      "step": 1871
+    },
+    {
+      "epoch": 0.9984,
+      "grad_norm": 0.36964428424835205,
+      "learning_rate": 3e-06,
+      "loss": 0.0771,
+      "policy/loss": 0.1562168262898922,
+      "step": 1872
+    },
+    {
+      "completion_length": 236.59375762939453,
+      "epoch": 1.0005333333333333,
+      "grad_norm": 0.28060758113861084,
+      "learning_rate": 3e-06,
+      "loss": -0.087,
+      "policy/loss": -0.1291867437558878,
+      "reward": 1.3802083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.6666666567325592,
+      "step": 1873
+    },
+    {
+      "epoch": 1.0010666666666668,
+      "grad_norm": 0.2242015302181244,
+      "learning_rate": 3e-06,
+      "loss": 0.0249,
+      "policy/loss": -0.01747538490572964,
+      "step": 1874
+    },
+    {
+      "epoch": 1.0016,
+      "grad_norm": 0.29876935482025146,
+      "learning_rate": 3e-06,
+      "loss": -0.0877,
+      "policy/loss": -0.13075114212463035,
+      "step": 1875
+    },
+    {
+      "epoch": 1.0021333333333333,
+      "grad_norm": 0.2631676495075226,
+      "learning_rate": 3e-06,
+      "loss": 0.0245,
+      "policy/loss": -0.016429857961647087,
+      "step": 1876
+    },
+    {
+      "completion_length": 221.06250762939453,
+      "epoch": 1.0026666666666666,
+      "grad_norm": 0.26835450530052185,
+      "learning_rate": 3e-06,
+      "loss": -0.1773,
+      "policy/loss": -0.44920861783734267,
+      "reward": 1.6093750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.8958333134651184,
+      "step": 1877
+    },
+    {
+      "epoch": 1.0032,
+      "grad_norm": 0.3228655755519867,
+      "learning_rate": 3e-06,
+      "loss": -0.2025,
+      "policy/loss": -0.3105650379083329,
+      "step": 1878
+    },
+    {
+      "epoch": 1.0037333333333334,
+      "grad_norm": 0.3249472379684448,
+      "learning_rate": 3e-06,
+      "loss": -0.1785,
+      "policy/loss": -0.44975659312552096,
+      "step": 1879
+    },
+    {
+      "epoch": 1.0042666666666666,
+      "grad_norm": 0.3235167860984802,
+      "learning_rate": 3e-06,
+      "loss": -0.2017,
+      "policy/loss": -0.3060232831942997,
+      "step": 1880
+    },
+    {
+      "completion_length": 211.7916717529297,
+      "epoch": 1.0048,
+      "grad_norm": 0.23878949880599976,
+      "learning_rate": 3e-06,
+      "loss": -0.1276,
+      "policy/loss": 1.0301475583673891e-07,
+      "reward": 1.3229166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.59375,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 1881
+    },
+    {
+      "epoch": 1.0053333333333334,
+      "grad_norm": 0.24519875645637512,
+      "learning_rate": 3e-06,
+      "loss": 0.2317,
+      "policy/loss": 8.333992340681107e-08,
+      "step": 1882
+    },
+    {
+      "epoch": 1.0058666666666667,
+      "grad_norm": 0.2391478717327118,
+      "learning_rate": 3e-06,
+      "loss": -0.1286,
+      "policy/loss": 1.0318476739712423e-07,
+      "step": 1883
+    },
+    {
+      "epoch": 1.0064,
+      "grad_norm": 0.2280661016702652,
+      "learning_rate": 3e-06,
+      "loss": 0.231,
+      "policy/loss": 8.329171663490342e-08,
+      "step": 1884
+    },
+    {
+      "completion_length": 228.00000762939453,
+      "epoch": 1.0069333333333332,
+      "grad_norm": 0.25946059823036194,
+      "learning_rate": 3e-06,
+      "loss": 0.1792,
+      "policy/loss": 7.704924342988306e-08,
+      "reward": 1.3229166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6041666716337204,
+      "step": 1885
+    },
+    {
+      "epoch": 1.0074666666666667,
+      "grad_norm": 0.2504805624485016,
+      "learning_rate": 3e-06,
+      "loss": 0.0109,
+      "policy/loss": 1.0593117849566624e-07,
+      "step": 1886
+    },
+    {
+      "epoch": 1.008,
+      "grad_norm": 0.30111631751060486,
+      "learning_rate": 3e-06,
+      "loss": 0.1785,
+      "policy/loss": 7.704838189681595e-08,
+      "step": 1887
+    },
+    {
+      "epoch": 1.0085333333333333,
+      "grad_norm": 0.25079697370529175,
+      "learning_rate": 3e-06,
+      "loss": 0.0102,
+      "policy/loss": 1.0597633703923748e-07,
+      "step": 1888
+    },
+    {
+      "completion_length": 233.03125,
+      "epoch": 1.0090666666666666,
+      "grad_norm": 0.2713717520236969,
+      "learning_rate": 3e-06,
+      "loss": -0.194,
+      "policy/loss": 0.014026288825906619,
+      "reward": 1.4583333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.7083333432674408,
+      "step": 1889
+    },
+    {
+      "epoch": 1.0096,
+      "grad_norm": 0.2745632529258728,
+      "learning_rate": 3e-06,
+      "loss": 0.0381,
+      "policy/loss": -0.10400009958393852,
+      "step": 1890
+    },
+    {
+      "epoch": 1.0101333333333333,
+      "grad_norm": 0.2850027084350586,
+      "learning_rate": 3e-06,
+      "loss": -0.1953,
+      "policy/loss": 0.012869879511757887,
+      "step": 1891
+    },
+    {
+      "epoch": 1.0106666666666666,
+      "grad_norm": 0.3312618136405945,
+      "learning_rate": 3e-06,
+      "loss": 0.036,
+      "policy/loss": -0.10846211073722856,
+      "step": 1892
+    },
+    {
+      "completion_length": 221.2604217529297,
+      "epoch": 1.0112,
+      "grad_norm": 0.21240785717964172,
+      "learning_rate": 3e-06,
+      "loss": 0.0192,
+      "policy/loss": -0.2335611537654252,
+      "reward": 1.1822916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.4374999850988388,
+      "step": 1893
+    },
+    {
+      "epoch": 1.0117333333333334,
+      "grad_norm": 0.33504757285118103,
+      "learning_rate": 3e-06,
+      "loss": 0.0811,
+      "policy/loss": -0.17816835597917446,
+      "step": 1894
+    },
+    {
+      "epoch": 1.0122666666666666,
+      "grad_norm": 0.2237490862607956,
+      "learning_rate": 3e-06,
+      "loss": 0.0189,
+      "policy/loss": -0.23193493034263213,
+      "step": 1895
+    },
+    {
+      "epoch": 1.0128,
+      "grad_norm": 0.35644781589508057,
+      "learning_rate": 3e-06,
+      "loss": 0.0795,
+      "policy/loss": -0.17776085394915953,
+      "step": 1896
+    },
+    {
+      "completion_length": 219.4791717529297,
+      "epoch": 1.0133333333333334,
+      "grad_norm": 0.33265775442123413,
+      "learning_rate": 3e-06,
+      "loss": -0.1278,
+      "policy/loss": 6.881175984574384e-08,
+      "reward": 1.625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 1.0,
+      "step": 1897
+    },
+    {
+      "epoch": 1.0138666666666667,
+      "grad_norm": 0.3061847388744354,
+      "learning_rate": 3e-06,
+      "loss": -0.1353,
+      "policy/loss": 9.04728452155723e-08,
+      "step": 1898
+    },
+    {
+      "epoch": 1.0144,
+      "grad_norm": 0.3156650960445404,
+      "learning_rate": 3e-06,
+      "loss": -0.1288,
+      "policy/loss": 6.885473169404577e-08,
+      "step": 1899
+    },
+    {
+      "epoch": 1.0149333333333332,
+      "grad_norm": 0.3036905527114868,
+      "learning_rate": 3e-06,
+      "loss": -0.1346,
+      "policy/loss": 9.043038318168328e-08,
+      "step": 1900
+    },
+    {
+      "completion_length": 214.87500762939453,
+      "epoch": 1.0154666666666667,
+      "grad_norm": 0.3159904181957245,
+      "learning_rate": 3e-06,
+      "loss": 0.0487,
+      "policy/loss": 1.3419791855540097e-07,
+      "reward": 1.1406250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.4583333432674408,
+      "step": 1901
+    },
+    {
+      "epoch": 1.016,
+      "grad_norm": 0.3303341269493103,
+      "learning_rate": 3e-06,
+      "loss": 0.1492,
+      "policy/loss": 1.784744014798889e-07,
+      "step": 1902
+    },
+    {
+      "epoch": 1.0165333333333333,
+      "grad_norm": 0.3110255002975464,
+      "learning_rate": 3e-06,
+      "loss": 0.0466,
+      "policy/loss": 1.3418229016792793e-07,
+      "step": 1903
+    },
+    {
+      "epoch": 1.0170666666666666,
+      "grad_norm": 0.311960905790329,
+      "learning_rate": 3e-06,
+      "loss": 0.1479,
+      "policy/loss": 1.7834602061839178e-07,
+      "step": 1904
+    },
+    {
+      "completion_length": 235.3229217529297,
+      "epoch": 1.0176,
+      "grad_norm": 0.2697281241416931,
+      "learning_rate": 3e-06,
+      "loss": -0.0115,
+      "policy/loss": -0.11871916032944796,
+      "reward": 1.3072916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.703125,
+      "rewards/correctness_reward_func_math": 0.6041666567325592,
+      "step": 1905
+    },
+    {
+      "epoch": 1.0181333333333333,
+      "grad_norm": 0.30777353048324585,
+      "learning_rate": 3e-06,
+      "loss": 0.0562,
+      "policy/loss": 0.29569107063435496,
+      "step": 1906
+    },
+    {
+      "epoch": 1.0186666666666666,
+      "grad_norm": 0.27213725447654724,
+      "learning_rate": 3e-06,
+      "loss": -0.0137,
+      "policy/loss": -0.1216826210257338,
+      "step": 1907
+    },
+    {
+      "epoch": 1.0192,
+      "grad_norm": 0.3012810945510864,
+      "learning_rate": 3e-06,
+      "loss": 0.0536,
+      "policy/loss": 0.2939105607595778,
+      "step": 1908
+    },
+    {
+      "completion_length": 223.28125,
+      "epoch": 1.0197333333333334,
+      "grad_norm": 0.32413360476493835,
+      "learning_rate": 3e-06,
+      "loss": -0.0422,
+      "policy/loss": -0.09222500243508591,
+      "reward": 1.40625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 1909
+    },
+    {
+      "epoch": 1.0202666666666667,
+      "grad_norm": 0.23187454044818878,
+      "learning_rate": 3e-06,
+      "loss": -0.1056,
+      "policy/loss": -0.04322043736011949,
+      "step": 1910
+    },
+    {
+      "epoch": 1.0208,
+      "grad_norm": 0.3013685941696167,
+      "learning_rate": 3e-06,
+      "loss": -0.0417,
+      "policy/loss": -0.08937734574067235,
+      "step": 1911
+    },
+    {
+      "epoch": 1.0213333333333334,
+      "grad_norm": 0.22933945059776306,
+      "learning_rate": 3e-06,
+      "loss": -0.1068,
+      "policy/loss": -0.04368588647612626,
+      "step": 1912
+    },
+    {
+      "completion_length": 209.5625,
+      "epoch": 1.0218666666666667,
+      "grad_norm": 0.2954249680042267,
+      "learning_rate": 3e-06,
+      "loss": -0.1175,
+      "policy/loss": 0.07323564432544316,
+      "reward": 1.4843750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.8333333134651184,
+      "step": 1913
+    },
+    {
+      "epoch": 1.0224,
+      "grad_norm": 0.24692191183567047,
+      "learning_rate": 3e-06,
+      "loss": 0.0476,
+      "policy/loss": 0.31166736311858756,
+      "step": 1914
+    },
+    {
+      "epoch": 1.0229333333333333,
+      "grad_norm": 0.2925046682357788,
+      "learning_rate": 3e-06,
+      "loss": -0.1184,
+      "policy/loss": 0.07134558098226762,
+      "step": 1915
+    },
+    {
+      "epoch": 1.0234666666666667,
+      "grad_norm": 0.2618328034877777,
+      "learning_rate": 3e-06,
+      "loss": 0.047,
+      "policy/loss": 0.3110690517092749,
+      "step": 1916
+    },
+    {
+      "completion_length": 225.28125762939453,
+      "epoch": 1.024,
+      "grad_norm": 0.27710849046707153,
+      "learning_rate": 3e-06,
+      "loss": -0.1192,
+      "policy/loss": 0.702002965995959,
+      "reward": 1.3750000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 1917
+    },
+    {
+      "epoch": 1.0245333333333333,
+      "grad_norm": 0.31901049613952637,
+      "learning_rate": 3e-06,
+      "loss": -0.1096,
+      "policy/loss": -0.4114521648597531,
+      "step": 1918
+    },
+    {
+      "epoch": 1.0250666666666666,
+      "grad_norm": 0.2737005650997162,
+      "learning_rate": 3e-06,
+      "loss": -0.1201,
+      "policy/loss": 0.6942376729438706,
+      "step": 1919
+    },
+    {
+      "epoch": 1.0256,
+      "grad_norm": 0.31783702969551086,
+      "learning_rate": 3e-06,
+      "loss": -0.1106,
+      "policy/loss": -0.4157541300465155,
+      "step": 1920
+    },
+    {
+      "completion_length": 227.90625762939453,
+      "epoch": 1.0261333333333333,
+      "grad_norm": 0.37542152404785156,
+      "learning_rate": 3e-06,
+      "loss": 0.1108,
+      "policy/loss": 8.577178789437312e-08,
+      "reward": 1.03125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.375,
+      "step": 1921
+    },
+    {
+      "epoch": 1.0266666666666666,
+      "grad_norm": 0.28465506434440613,
+      "learning_rate": 3e-06,
+      "loss": 0.0719,
+      "policy/loss": 7.601536822221533e-08,
+      "step": 1922
+    },
+    {
+      "epoch": 1.0272,
+      "grad_norm": 0.22451888024806976,
+      "learning_rate": 3e-06,
+      "loss": 0.1112,
+      "policy/loss": 8.576628474088466e-08,
+      "step": 1923
+    },
+    {
+      "epoch": 1.0277333333333334,
+      "grad_norm": 0.24187196791172028,
+      "learning_rate": 3e-06,
+      "loss": 0.0698,
+      "policy/loss": 7.608207752696217e-08,
+      "step": 1924
+    },
+    {
+      "completion_length": 229.5416717529297,
+      "epoch": 1.0282666666666667,
+      "grad_norm": 0.2995518147945404,
+      "learning_rate": 3e-06,
+      "loss": -0.0396,
+      "policy/loss": -0.0718383640050888,
+      "reward": 1.5,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 1925
+    },
+    {
+      "epoch": 1.0288,
+      "grad_norm": 0.3647313714027405,
+      "learning_rate": 3e-06,
+      "loss": 0.0125,
+      "policy/loss": -0.48488805443048477,
+      "step": 1926
+    },
+    {
+      "epoch": 1.0293333333333334,
+      "grad_norm": 0.3125527799129486,
+      "learning_rate": 3e-06,
+      "loss": -0.042,
+      "policy/loss": -0.07763822376728058,
+      "step": 1927
+    },
+    {
+      "epoch": 1.0298666666666667,
+      "grad_norm": 0.3369147479534149,
+      "learning_rate": 3e-06,
+      "loss": 0.0115,
+      "policy/loss": -0.4855204448103905,
+      "step": 1928
+    },
+    {
+      "completion_length": 214.40625762939453,
+      "epoch": 1.0304,
+      "grad_norm": 0.3307705819606781,
+      "learning_rate": 3e-06,
+      "loss": 0.246,
+      "policy/loss": 0.019664607161534065,
+      "reward": 1.5156250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6197916567325592,
+      "rewards/correctness_reward_func_math": 0.8958333134651184,
+      "step": 1929
+    },
+    {
+      "epoch": 1.0309333333333333,
+      "grad_norm": 0.4678812325000763,
+      "learning_rate": 3e-06,
+      "loss": 0.1295,
+      "policy/loss": 0.09453501169411638,
+      "step": 1930
+    },
+    {
+      "epoch": 1.0314666666666668,
+      "grad_norm": 0.28450536727905273,
+      "learning_rate": 3e-06,
+      "loss": 0.245,
+      "policy/loss": 0.019686519369003008,
+      "step": 1931
+    },
+    {
+      "epoch": 1.032,
+      "grad_norm": 0.3928939402103424,
+      "learning_rate": 3e-06,
+      "loss": 0.1264,
+      "policy/loss": 0.09470291053710866,
+      "step": 1932
+    },
+    {
+      "completion_length": 223.40625762939453,
+      "epoch": 1.0325333333333333,
+      "grad_norm": 0.27206408977508545,
+      "learning_rate": 3e-06,
+      "loss": -0.0153,
+      "policy/loss": 0.32908103056252,
+      "reward": 1.5989583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.9166666567325592,
+      "step": 1933
+    },
+    {
+      "epoch": 1.0330666666666666,
+      "grad_norm": 0.24129608273506165,
+      "learning_rate": 3e-06,
+      "loss": 0.113,
+      "policy/loss": 0.7853658003732562,
+      "step": 1934
+    },
+    {
+      "epoch": 1.0336,
+      "grad_norm": 0.6620590090751648,
+      "learning_rate": 3e-06,
+      "loss": -0.0146,
+      "policy/loss": 0.3285087216645479,
+      "step": 1935
+    },
+    {
+      "epoch": 1.0341333333333333,
+      "grad_norm": 0.23715968430042267,
+      "learning_rate": 3e-06,
+      "loss": 0.1135,
+      "policy/loss": 0.7857936853542924,
+      "step": 1936
+    },
+    {
+      "completion_length": 236.95833587646484,
+      "epoch": 1.0346666666666666,
+      "grad_norm": 0.28248852491378784,
+      "learning_rate": 3e-06,
+      "loss": -0.0041,
+      "policy/loss": 0.4123106002807617,
+      "reward": 1.0572916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.375,
+      "step": 1937
+    },
+    {
+      "epoch": 1.0352,
+      "grad_norm": 0.3132759928703308,
+      "learning_rate": 3e-06,
+      "loss": -0.2056,
+      "policy/loss": -0.49341627955436707,
+      "step": 1938
+    },
+    {
+      "epoch": 1.0357333333333334,
+      "grad_norm": 0.28405463695526123,
+      "learning_rate": 3e-06,
+      "loss": -0.0046,
+      "policy/loss": 0.41068539023399353,
+      "step": 1939
+    },
+    {
+      "epoch": 1.0362666666666667,
+      "grad_norm": 0.39635759592056274,
+      "learning_rate": 3e-06,
+      "loss": -0.2071,
+      "policy/loss": -0.4997134506702423,
+      "step": 1940
+    },
+    {
+      "completion_length": 236.21875762939453,
+      "epoch": 1.0368,
+      "grad_norm": 0.20278042554855347,
+      "learning_rate": 3e-06,
+      "loss": -0.0364,
+      "policy/loss": -0.4355386048555374,
+      "reward": 1.5104166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6979166567325592,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 1941
+    },
+    {
+      "epoch": 1.0373333333333334,
+      "grad_norm": 0.21029607951641083,
+      "learning_rate": 3e-06,
+      "loss": 0.0463,
+      "policy/loss": 0.18448693864047527,
+      "step": 1942
+    },
+    {
+      "epoch": 1.0378666666666667,
+      "grad_norm": 0.19959059357643127,
+      "learning_rate": 3e-06,
+      "loss": -0.0381,
+      "policy/loss": -0.43886326253414154,
+      "step": 1943
+    },
+    {
+      "epoch": 1.0384,
+      "grad_norm": 0.20707404613494873,
+      "learning_rate": 3e-06,
+      "loss": 0.0453,
+      "policy/loss": 0.18186903931200504,
+      "step": 1944
+    },
+    {
+      "completion_length": 213.5104217529297,
+      "epoch": 1.0389333333333333,
+      "grad_norm": 0.2243618369102478,
+      "learning_rate": 3e-06,
+      "loss": 0.0742,
+      "policy/loss": 0.20971041172742844,
+      "reward": 1.6145833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8958333432674408,
+      "step": 1945
+    },
+    {
+      "epoch": 1.0394666666666668,
+      "grad_norm": 0.19831636548042297,
+      "learning_rate": 3e-06,
+      "loss": 0.1158,
+      "policy/loss": 0.4554571956396103,
+      "step": 1946
+    },
+    {
+      "epoch": 1.04,
+      "grad_norm": 0.26221299171447754,
+      "learning_rate": 3e-06,
+      "loss": 0.0738,
+      "policy/loss": 0.20944903418421745,
+      "step": 1947
+    },
+    {
+      "epoch": 1.0405333333333333,
+      "grad_norm": 0.1951523721218109,
+      "learning_rate": 3e-06,
+      "loss": 0.1145,
+      "policy/loss": 0.45333556830883026,
+      "step": 1948
+    },
+    {
+      "completion_length": 232.18750762939453,
+      "epoch": 1.0410666666666666,
+      "grad_norm": 0.3079296946525574,
+      "learning_rate": 3e-06,
+      "loss": 0.1151,
+      "policy/loss": -0.14944683894506028,
+      "reward": 1.2395833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.5208333432674408,
+      "step": 1949
+    },
+    {
+      "epoch": 1.0416,
+      "grad_norm": 0.2929827570915222,
+      "learning_rate": 3e-06,
+      "loss": 0.0837,
+      "policy/loss": 0.3858074015947892,
+      "step": 1950
+    },
+    {
+      "epoch": 1.0421333333333334,
+      "grad_norm": 0.3120653033256531,
+      "learning_rate": 3e-06,
+      "loss": 0.114,
+      "policy/loss": -0.1516347466256498,
+      "step": 1951
+    },
+    {
+      "epoch": 1.0426666666666666,
+      "grad_norm": 0.3102172315120697,
+      "learning_rate": 3e-06,
+      "loss": 0.0825,
+      "policy/loss": 0.3839093809712679,
+      "step": 1952
+    },
+    {
+      "completion_length": 244.5729217529297,
+      "epoch": 1.0432,
+      "grad_norm": 0.2530492842197418,
+      "learning_rate": 3e-06,
+      "loss": 0.0007,
+      "policy/loss": 1.0341208778186228e-07,
+      "reward": 1.1354166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6770833134651184,
+      "rewards/correctness_reward_func_math": 0.4583333283662796,
+      "step": 1953
+    },
+    {
+      "epoch": 1.0437333333333334,
+      "grad_norm": 0.7409317493438721,
+      "learning_rate": 3e-06,
+      "loss": -0.0388,
+      "policy/loss": 1.1298091884270889e-07,
+      "step": 1954
+    },
+    {
+      "epoch": 1.0442666666666667,
+      "grad_norm": 0.24981433153152466,
+      "learning_rate": 3e-06,
+      "loss": -0.0009,
+      "policy/loss": 1.034094960772336e-07,
+      "step": 1955
+    },
+    {
+      "epoch": 1.0448,
+      "grad_norm": 0.3463074564933777,
+      "learning_rate": 3e-06,
+      "loss": -0.0459,
+      "policy/loss": 1.1312011949371481e-07,
+      "step": 1956
+    },
+    {
+      "completion_length": 230.4166717529297,
+      "epoch": 1.0453333333333332,
+      "grad_norm": 0.13285893201828003,
+      "learning_rate": 3e-06,
+      "loss": -0.1089,
+      "policy/loss": 5.8132908264951766e-08,
+      "reward": 1.1458333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.4583333283662796,
+      "step": 1957
+    },
+    {
+      "epoch": 1.0458666666666667,
+      "grad_norm": 0.1695667952299118,
+      "learning_rate": 3e-06,
+      "loss": 0.0403,
+      "policy/loss": 6.598031632165657e-08,
+      "step": 1958
+    },
+    {
+      "epoch": 1.0464,
+      "grad_norm": 0.13321730494499207,
+      "learning_rate": 3e-06,
+      "loss": -0.1091,
+      "policy/loss": 5.807821779058031e-08,
+      "step": 1959
+    },
+    {
+      "epoch": 1.0469333333333333,
+      "grad_norm": 0.1981400102376938,
+      "learning_rate": 3e-06,
+      "loss": 0.0397,
+      "policy/loss": 6.60089867210445e-08,
+      "step": 1960
+    },
+    {
+      "completion_length": 230.5729217529297,
+      "epoch": 1.0474666666666668,
+      "grad_norm": 0.566965639591217,
+      "learning_rate": 3e-06,
+      "loss": 0.0855,
+      "policy/loss": 6.101768956767728e-08,
+      "reward": 1.6770833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.9583333134651184,
+      "step": 1961
+    },
+    {
+      "epoch": 1.048,
+      "grad_norm": 0.2915070354938507,
+      "learning_rate": 3e-06,
+      "loss": 0.002,
+      "policy/loss": 7.165472837300513e-08,
+      "step": 1962
+    },
+    {
+      "epoch": 1.0485333333333333,
+      "grad_norm": 0.540489673614502,
+      "learning_rate": 3e-06,
+      "loss": 0.0846,
+      "policy/loss": 6.110150607696596e-08,
+      "step": 1963
+    },
+    {
+      "epoch": 1.0490666666666666,
+      "grad_norm": 0.2666778266429901,
+      "learning_rate": 3e-06,
+      "loss": 0.0011,
+      "policy/loss": 7.180665129169483e-08,
+      "step": 1964
+    },
+    {
+      "completion_length": 228.73959350585938,
+      "epoch": 1.0496,
+      "grad_norm": 0.4486520290374756,
+      "learning_rate": 3e-06,
+      "loss": 0.0691,
+      "policy/loss": 0.4256247729063034,
+      "reward": 1.5833333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.8333333134651184,
+      "step": 1965
+    },
+    {
+      "epoch": 1.0501333333333334,
+      "grad_norm": 0.32992610335350037,
+      "learning_rate": 3e-06,
+      "loss": 0.2444,
+      "policy/loss": 0.5406920313835144,
+      "step": 1966
+    },
+    {
+      "epoch": 1.0506666666666666,
+      "grad_norm": 0.4141113758087158,
+      "learning_rate": 3e-06,
+      "loss": 0.0665,
+      "policy/loss": 0.41641464829444885,
+      "step": 1967
+    },
+    {
+      "epoch": 1.0512,
+      "grad_norm": 0.3458501994609833,
+      "learning_rate": 3e-06,
+      "loss": 0.2433,
+      "policy/loss": 0.5383584946393967,
+      "step": 1968
+    },
+    {
+      "completion_length": 220.87500762939453,
+      "epoch": 1.0517333333333334,
+      "grad_norm": 0.2178836613893509,
+      "learning_rate": 3e-06,
+      "loss": 0.04,
+      "policy/loss": -0.04162070155143738,
+      "reward": 1.5104166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 1969
+    },
+    {
+      "epoch": 1.0522666666666667,
+      "grad_norm": 0.2731064260005951,
+      "learning_rate": 3e-06,
+      "loss": -0.0109,
+      "policy/loss": -0.00022376360720954835,
+      "step": 1970
+    },
+    {
+      "epoch": 1.0528,
+      "grad_norm": 0.2228037267923355,
+      "learning_rate": 3e-06,
+      "loss": 0.0391,
+      "policy/loss": -0.04194958880543709,
+      "step": 1971
+    },
+    {
+      "epoch": 1.0533333333333332,
+      "grad_norm": 0.2804855406284332,
+      "learning_rate": 3e-06,
+      "loss": -0.0111,
+      "policy/loss": 8.361768414033577e-05,
+      "step": 1972
+    },
+    {
+      "completion_length": 240.30208587646484,
+      "epoch": 1.0538666666666667,
+      "grad_norm": 0.24198703467845917,
+      "learning_rate": 3e-06,
+      "loss": 0.1418,
+      "policy/loss": 0.6065249443054199,
+      "reward": 1.109375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5885416716337204,
+      "rewards/correctness_reward_func_math": 0.5208333283662796,
+      "step": 1973
+    },
+    {
+      "epoch": 1.0544,
+      "grad_norm": 0.26036444306373596,
+      "learning_rate": 3e-06,
+      "loss": 0.0845,
+      "policy/loss": 0.16613066382706165,
+      "step": 1974
+    },
+    {
+      "epoch": 1.0549333333333333,
+      "grad_norm": 0.24229079484939575,
+      "learning_rate": 3e-06,
+      "loss": 0.1411,
+      "policy/loss": 0.6065394282341003,
+      "step": 1975
+    },
+    {
+      "epoch": 1.0554666666666668,
+      "grad_norm": 0.2175290435552597,
+      "learning_rate": 3e-06,
+      "loss": 0.0836,
+      "policy/loss": 0.16095904260873795,
+      "step": 1976
+    },
+    {
+      "completion_length": 218.1666717529297,
+      "epoch": 1.056,
+      "grad_norm": 0.30513548851013184,
+      "learning_rate": 3e-06,
+      "loss": -0.1266,
+      "policy/loss": -0.7401886381430938,
+      "reward": 1.7083333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 1.0208333432674408,
+      "step": 1977
+    },
+    {
+      "epoch": 1.0565333333333333,
+      "grad_norm": 0.22399315237998962,
+      "learning_rate": 3e-06,
+      "loss": -0.0382,
+      "policy/loss": 0.1618777257184938,
+      "step": 1978
+    },
+    {
+      "epoch": 1.0570666666666666,
+      "grad_norm": 0.3014119565486908,
+      "learning_rate": 3e-06,
+      "loss": -0.1271,
+      "policy/loss": -0.7413623727933629,
+      "step": 1979
+    },
+    {
+      "epoch": 1.0576,
+      "grad_norm": 0.22060935199260712,
+      "learning_rate": 3e-06,
+      "loss": -0.0394,
+      "policy/loss": 0.16128954712838706,
+      "step": 1980
+    },
+    {
+      "completion_length": 240.73958587646484,
+      "epoch": 1.0581333333333334,
+      "grad_norm": 0.9425589442253113,
+      "learning_rate": 3e-06,
+      "loss": -0.0019,
+      "policy/loss": 1.200277495172486e-07,
+      "reward": 1.2447916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.5000000149011612,
+      "step": 1981
+    },
+    {
+      "epoch": 1.0586666666666666,
+      "grad_norm": 0.3661278784275055,
+      "learning_rate": 3e-06,
+      "loss": -0.144,
+      "policy/loss": 1.0541976180888923e-07,
+      "step": 1982
+    },
+    {
+      "epoch": 1.0592,
+      "grad_norm": 0.8086930513381958,
+      "learning_rate": 3e-06,
+      "loss": -0.0032,
+      "policy/loss": 1.204776474139635e-07,
+      "step": 1983
+    },
+    {
+      "epoch": 1.0597333333333334,
+      "grad_norm": 0.38221287727355957,
+      "learning_rate": 3e-06,
+      "loss": -0.1458,
+      "policy/loss": 1.0539959305333468e-07,
+      "step": 1984
+    },
+    {
+      "completion_length": 229.6666717529297,
+      "epoch": 1.0602666666666667,
+      "grad_norm": 0.29671433568000793,
+      "learning_rate": 3e-06,
+      "loss": 0.0592,
+      "policy/loss": 0.20055639159936334,
+      "reward": 1.5,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 1985
+    },
+    {
+      "epoch": 1.0608,
+      "grad_norm": 0.31810030341148376,
+      "learning_rate": 3e-06,
+      "loss": 0.051,
+      "policy/loss": 0.2961098595875562,
+      "step": 1986
+    },
+    {
+      "epoch": 1.0613333333333332,
+      "grad_norm": 0.2749922573566437,
+      "learning_rate": 3e-06,
+      "loss": 0.058,
+      "policy/loss": 0.19448798316004812,
+      "step": 1987
+    },
+    {
+      "epoch": 1.0618666666666667,
+      "grad_norm": 0.32830485701560974,
+      "learning_rate": 3e-06,
+      "loss": 0.05,
+      "policy/loss": 0.2931880399301363,
+      "step": 1988
+    },
+    {
+      "completion_length": 225.8125,
+      "epoch": 1.0624,
+      "grad_norm": 0.24826055765151978,
+      "learning_rate": 3e-06,
+      "loss": 0.1181,
+      "policy/loss": 1.0956213891688549e-07,
+      "reward": 1.09375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.4375,
+      "step": 1989
+    },
+    {
+      "epoch": 1.0629333333333333,
+      "grad_norm": 0.26565733551979065,
+      "learning_rate": 3e-06,
+      "loss": 0.032,
+      "policy/loss": 1.282461923324263e-07,
+      "step": 1990
+    },
+    {
+      "epoch": 1.0634666666666668,
+      "grad_norm": 0.25404807925224304,
+      "learning_rate": 3e-06,
+      "loss": 0.1171,
+      "policy/loss": 1.0956880380774692e-07,
+      "step": 1991
+    },
+    {
+      "epoch": 1.064,
+      "grad_norm": 0.28134170174598694,
+      "learning_rate": 3e-06,
+      "loss": 0.0291,
+      "policy/loss": 1.2816973082863115e-07,
+      "step": 1992
+    },
+    {
+      "completion_length": 241.15625762939453,
+      "epoch": 1.0645333333333333,
+      "grad_norm": 0.36316242814064026,
+      "learning_rate": 3e-06,
+      "loss": -0.0084,
+      "policy/loss": 0.35394250229001045,
+      "reward": 1.1979167461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.4791666716337204,
+      "step": 1993
+    },
+    {
+      "epoch": 1.0650666666666666,
+      "grad_norm": 0.41210120916366577,
+      "learning_rate": 3e-06,
+      "loss": -0.171,
+      "policy/loss": -0.7859403491020203,
+      "step": 1994
+    },
+    {
+      "epoch": 1.0656,
+      "grad_norm": 0.34867992997169495,
+      "learning_rate": 3e-06,
+      "loss": -0.0093,
+      "policy/loss": 0.35154804959893227,
+      "step": 1995
+    },
+    {
+      "epoch": 1.0661333333333334,
+      "grad_norm": 0.38860756158828735,
+      "learning_rate": 3e-06,
+      "loss": -0.1712,
+      "policy/loss": -0.790502741932869,
+      "step": 1996
+    },
+    {
+      "completion_length": 221.02083587646484,
+      "epoch": 1.0666666666666667,
+      "grad_norm": 0.2336549162864685,
+      "learning_rate": 3e-06,
+      "loss": 0.0313,
+      "policy/loss": 1.1828014478965088e-07,
+      "reward": 1.6093750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.9583333432674408,
+      "step": 1997
+    },
+    {
+      "epoch": 1.0672,
+      "grad_norm": 0.24817509949207306,
+      "learning_rate": 3e-06,
+      "loss": -0.1589,
+      "policy/loss": 1.2653619307911868e-07,
+      "step": 1998
+    },
+    {
+      "epoch": 1.0677333333333334,
+      "grad_norm": 0.2788016200065613,
+      "learning_rate": 3e-06,
+      "loss": 0.0291,
+      "policy/loss": 1.1837815705462162e-07,
+      "step": 1999
+    },
+    {
+      "epoch": 1.0682666666666667,
+      "grad_norm": 0.26208987832069397,
+      "learning_rate": 3e-06,
+      "loss": -0.1604,
+      "policy/loss": 1.2651445047140442e-07,
+      "step": 2000
+    },
+    {
+      "completion_length": 231.40625762939453,
+      "epoch": 1.0688,
+      "grad_norm": 0.21465539932250977,
+      "learning_rate": 3e-06,
+      "loss": -0.1476,
+      "policy/loss": -0.6372177978607567,
+      "reward": 1.5312500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.875,
+      "step": 2001
+    },
+    {
+      "epoch": 1.0693333333333332,
+      "grad_norm": 0.20916689932346344,
+      "learning_rate": 3e-06,
+      "loss": -0.0295,
+      "policy/loss": -0.010876083750289922,
+      "step": 2002
+    },
+    {
+      "epoch": 1.0698666666666667,
+      "grad_norm": 0.20814698934555054,
+      "learning_rate": 3e-06,
+      "loss": -0.1483,
+      "policy/loss": -0.6360483547131679,
+      "step": 2003
+    },
+    {
+      "epoch": 1.0704,
+      "grad_norm": 0.1954680234193802,
+      "learning_rate": 3e-06,
+      "loss": -0.0301,
+      "policy/loss": -0.013412977597795361,
+      "step": 2004
+    },
+    {
+      "completion_length": 221.86458587646484,
+      "epoch": 1.0709333333333333,
+      "grad_norm": 0.42735642194747925,
+      "learning_rate": 3e-06,
+      "loss": -0.0291,
+      "policy/loss": 0.2566667540953489,
+      "reward": 1.2447916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.5625,
+      "step": 2005
+    },
+    {
+      "epoch": 1.0714666666666666,
+      "grad_norm": 0.23398925364017487,
+      "learning_rate": 3e-06,
+      "loss": -0.0018,
+      "policy/loss": 0.5447205602607088,
+      "step": 2006
+    },
+    {
+      "epoch": 1.072,
+      "grad_norm": 0.2512841522693634,
+      "learning_rate": 3e-06,
+      "loss": -0.0323,
+      "policy/loss": 0.2501077993584211,
+      "step": 2007
+    },
+    {
+      "epoch": 1.0725333333333333,
+      "grad_norm": 0.23921740055084229,
+      "learning_rate": 3e-06,
+      "loss": -0.0037,
+      "policy/loss": 0.532791525030575,
+      "step": 2008
+    },
+    {
+      "completion_length": 217.3229217529297,
+      "epoch": 1.0730666666666666,
+      "grad_norm": 0.3679647445678711,
+      "learning_rate": 3e-06,
+      "loss": -0.0457,
+      "policy/loss": -0.20342646911740303,
+      "reward": 1.7291666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.9791666567325592,
+      "step": 2009
+    },
+    {
+      "epoch": 1.0735999999999999,
+      "grad_norm": 1.0285983085632324,
+      "learning_rate": 3e-06,
+      "loss": 0.0065,
+      "policy/loss": 0.26556916534900665,
+      "step": 2010
+    },
+    {
+      "epoch": 1.0741333333333334,
+      "grad_norm": 0.36113402247428894,
+      "learning_rate": 3e-06,
+      "loss": -0.0479,
+      "policy/loss": -0.2137500774115324,
+      "step": 2011
+    },
+    {
+      "epoch": 1.0746666666666667,
+      "grad_norm": 1.4094462394714355,
+      "learning_rate": 3e-06,
+      "loss": 0.0041,
+      "policy/loss": 0.2650686949491501,
+      "step": 2012
+    },
+    {
+      "completion_length": 234.5,
+      "epoch": 1.0752,
+      "grad_norm": 0.2518153190612793,
+      "learning_rate": 3e-06,
+      "loss": -0.1978,
+      "policy/loss": 7.264648083094016e-08,
+      "reward": 1.7604166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 1.0416666865348816,
+      "step": 2013
+    },
+    {
+      "epoch": 1.0757333333333334,
+      "grad_norm": 0.28222572803497314,
+      "learning_rate": 3e-06,
+      "loss": -0.0486,
+      "policy/loss": 9.822057656094785e-08,
+      "step": 2014
+    },
+    {
+      "epoch": 1.0762666666666667,
+      "grad_norm": 0.2607128322124481,
+      "learning_rate": 3e-06,
+      "loss": -0.1987,
+      "policy/loss": 7.282588043722171e-08,
+      "step": 2015
+    },
+    {
+      "epoch": 1.0768,
+      "grad_norm": 0.27392858266830444,
+      "learning_rate": 3e-06,
+      "loss": -0.0485,
+      "policy/loss": 9.818958091045715e-08,
+      "step": 2016
+    },
+    {
+      "completion_length": 221.98958587646484,
+      "epoch": 1.0773333333333333,
+      "grad_norm": 0.33134499192237854,
+      "learning_rate": 3e-06,
+      "loss": -0.193,
+      "policy/loss": 0.06848351564258337,
+      "reward": 1.6302083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6927083432674408,
+      "rewards/correctness_reward_func_math": 0.9375,
+      "step": 2017
+    },
+    {
+      "epoch": 1.0778666666666668,
+      "grad_norm": 0.2856132388114929,
+      "learning_rate": 3e-06,
+      "loss": -0.0805,
+      "policy/loss": -0.06979334726929665,
+      "step": 2018
+    },
+    {
+      "epoch": 1.0784,
+      "grad_norm": 0.333514928817749,
+      "learning_rate": 3e-06,
+      "loss": -0.1942,
+      "policy/loss": 0.07042689900845289,
+      "step": 2019
+    },
+    {
+      "epoch": 1.0789333333333333,
+      "grad_norm": 0.2849968671798706,
+      "learning_rate": 3e-06,
+      "loss": -0.083,
+      "policy/loss": -0.07574868574738503,
+      "step": 2020
+    },
+    {
+      "completion_length": 236.18750762939453,
+      "epoch": 1.0794666666666666,
+      "grad_norm": 0.2281435877084732,
+      "learning_rate": 3e-06,
+      "loss": 0.0211,
+      "policy/loss": 0.07243780304036562,
+      "reward": 1.125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.4375,
+      "step": 2021
+    },
+    {
+      "epoch": 1.08,
+      "grad_norm": 0.27494508028030396,
+      "learning_rate": 3e-06,
+      "loss": 0.0779,
+      "policy/loss": -0.3143629650660884,
+      "step": 2022
+    },
+    {
+      "epoch": 1.0805333333333333,
+      "grad_norm": 0.22964803874492645,
+      "learning_rate": 3e-06,
+      "loss": 0.0206,
+      "policy/loss": 0.0758088854604182,
+      "step": 2023
+    },
+    {
+      "epoch": 1.0810666666666666,
+      "grad_norm": 0.28103107213974,
+      "learning_rate": 3e-06,
+      "loss": 0.0751,
+      "policy/loss": -0.3161997119168163,
+      "step": 2024
+    },
+    {
+      "completion_length": 220.70833587646484,
+      "epoch": 1.0816,
+      "grad_norm": 0.3107304275035858,
+      "learning_rate": 3e-06,
+      "loss": -0.1335,
+      "policy/loss": 0.11967246450060642,
+      "reward": 1.5729166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.9166666865348816,
+      "step": 2025
+    },
+    {
+      "epoch": 1.0821333333333334,
+      "grad_norm": 0.26591652631759644,
+      "learning_rate": 3e-06,
+      "loss": 0.0054,
+      "policy/loss": 0.02139433782863165,
+      "step": 2026
+    },
+    {
+      "epoch": 1.0826666666666667,
+      "grad_norm": 0.3187543749809265,
+      "learning_rate": 3e-06,
+      "loss": -0.1346,
+      "policy/loss": 0.1222524663894724,
+      "step": 2027
+    },
+    {
+      "epoch": 1.0832,
+      "grad_norm": 0.3286965489387512,
+      "learning_rate": 3e-06,
+      "loss": 0.0047,
+      "policy/loss": 0.02003760218732964,
+      "step": 2028
+    },
+    {
+      "completion_length": 220.96875762939453,
+      "epoch": 1.0837333333333334,
+      "grad_norm": 0.2710818648338318,
+      "learning_rate": 3e-06,
+      "loss": -0.0786,
+      "policy/loss": -0.28528093360364437,
+      "reward": 1.8593751192092896,
+      "rewards/boxed_and_answer_tags_format_reward": 0.734375,
+      "rewards/correctness_reward_func_math": 1.1249999701976776,
+      "step": 2029
+    },
+    {
+      "epoch": 1.0842666666666667,
+      "grad_norm": 0.25109267234802246,
+      "learning_rate": 3e-06,
+      "loss": -0.0386,
+      "policy/loss": -0.30708301067352295,
+      "step": 2030
+    },
+    {
+      "epoch": 1.0848,
+      "grad_norm": 0.2957793176174164,
+      "learning_rate": 3e-06,
+      "loss": -0.0793,
+      "policy/loss": -0.28665725886821747,
+      "step": 2031
+    },
+    {
+      "epoch": 1.0853333333333333,
+      "grad_norm": 0.24931909143924713,
+      "learning_rate": 3e-06,
+      "loss": -0.0403,
+      "policy/loss": -0.3136259615421295,
+      "step": 2032
+    },
+    {
+      "completion_length": 216.6666717529297,
+      "epoch": 1.0858666666666668,
+      "grad_norm": 0.2609615623950958,
+      "learning_rate": 3e-06,
+      "loss": 0.0211,
+      "policy/loss": 1.3592587322364125e-07,
+      "reward": 1.015625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5989583432674408,
+      "rewards/correctness_reward_func_math": 0.4166666679084301,
+      "step": 2033
+    },
+    {
+      "epoch": 1.0864,
+      "grad_norm": 0.23326577246189117,
+      "learning_rate": 3e-06,
+      "loss": -0.057,
+      "policy/loss": 1.7779542105245127e-07,
+      "step": 2034
+    },
+    {
+      "epoch": 1.0869333333333333,
+      "grad_norm": 0.2663092017173767,
+      "learning_rate": 3e-06,
+      "loss": 0.0203,
+      "policy/loss": 1.3582014801727382e-07,
+      "step": 2035
+    },
+    {
+      "epoch": 1.0874666666666666,
+      "grad_norm": 0.23955470323562622,
+      "learning_rate": 3e-06,
+      "loss": -0.0578,
+      "policy/loss": 1.7770808113937164e-07,
+      "step": 2036
+    },
+    {
+      "completion_length": 231.5,
+      "epoch": 1.088,
+      "grad_norm": 0.3177078068256378,
+      "learning_rate": 3e-06,
+      "loss": -0.1196,
+      "policy/loss": -0.2036406826718391,
+      "reward": 1.2395833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.5208333432674408,
+      "step": 2037
+    },
+    {
+      "epoch": 1.0885333333333334,
+      "grad_norm": 0.2601716220378876,
+      "learning_rate": 3e-06,
+      "loss": -0.0055,
+      "policy/loss": 0.12323352868570225,
+      "step": 2038
+    },
+    {
+      "epoch": 1.0890666666666666,
+      "grad_norm": 0.274558961391449,
+      "learning_rate": 3e-06,
+      "loss": -0.1189,
+      "policy/loss": -0.21054810104043753,
+      "step": 2039
+    },
+    {
+      "epoch": 1.0896,
+      "grad_norm": 0.27108368277549744,
+      "learning_rate": 3e-06,
+      "loss": -0.0063,
+      "policy/loss": 0.11852938854865513,
+      "step": 2040
+    },
+    {
+      "completion_length": 226.96875762939453,
+      "epoch": 1.0901333333333334,
+      "grad_norm": 0.33603209257125854,
+      "learning_rate": 3e-06,
+      "loss": -0.0156,
+      "policy/loss": -0.03093681320221009,
+      "reward": 1.7500000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 1.0624999701976776,
+      "step": 2041
+    },
+    {
+      "epoch": 1.0906666666666667,
+      "grad_norm": 0.2910367548465729,
+      "learning_rate": 3e-06,
+      "loss": -0.0741,
+      "policy/loss": 0.10459830392061065,
+      "step": 2042
+    },
+    {
+      "epoch": 1.0912,
+      "grad_norm": 0.3411917984485626,
+      "learning_rate": 3e-06,
+      "loss": -0.0171,
+      "policy/loss": -0.031922363061994474,
+      "step": 2043
+    },
+    {
+      "epoch": 1.0917333333333334,
+      "grad_norm": 0.2894488275051117,
+      "learning_rate": 3e-06,
+      "loss": -0.0761,
+      "policy/loss": 0.10320594687461693,
+      "step": 2044
+    },
+    {
+      "completion_length": 215.53125762939453,
+      "epoch": 1.0922666666666667,
+      "grad_norm": 0.2639469802379608,
+      "learning_rate": 3e-06,
+      "loss": -0.0409,
+      "policy/loss": 5.954476911540496e-08,
+      "reward": 1.5625000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.875,
+      "step": 2045
+    },
+    {
+      "epoch": 1.0928,
+      "grad_norm": 0.29702869057655334,
+      "learning_rate": 3e-06,
+      "loss": -0.1285,
+      "policy/loss": 6.754208392578676e-08,
+      "step": 2046
+    },
+    {
+      "epoch": 1.0933333333333333,
+      "grad_norm": 0.2605637311935425,
+      "learning_rate": 3e-06,
+      "loss": -0.0418,
+      "policy/loss": 5.960540150340421e-08,
+      "step": 2047
+    },
+    {
+      "epoch": 1.0938666666666668,
+      "grad_norm": 0.2912304699420929,
+      "learning_rate": 3e-06,
+      "loss": -0.13,
+      "policy/loss": 6.749320746735066e-08,
+      "step": 2048
+    },
+    {
+      "completion_length": 221.03125762939453,
+      "epoch": 1.0944,
+      "grad_norm": 0.28811362385749817,
+      "learning_rate": 3e-06,
+      "loss": 0.1352,
+      "policy/loss": 0.08815735297446814,
+      "reward": 1.1718750298023224,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.5208333246409893,
+      "step": 2049
+    },
+    {
+      "epoch": 1.0949333333333333,
+      "grad_norm": 0.29913973808288574,
+      "learning_rate": 3e-06,
+      "loss": -0.1237,
+      "policy/loss": -0.4221285438600857,
+      "step": 2050
+    },
+    {
+      "epoch": 1.0954666666666666,
+      "grad_norm": 0.3008749485015869,
+      "learning_rate": 3e-06,
+      "loss": 0.135,
+      "policy/loss": 0.08889008031392187,
+      "step": 2051
+    },
+    {
+      "epoch": 1.096,
+      "grad_norm": 0.3187890946865082,
+      "learning_rate": 3e-06,
+      "loss": -0.1254,
+      "policy/loss": -0.42460154057272614,
+      "step": 2052
+    },
+    {
+      "completion_length": 223.81250762939453,
+      "epoch": 1.0965333333333334,
+      "grad_norm": 0.29849398136138916,
+      "learning_rate": 3e-06,
+      "loss": -0.0039,
+      "policy/loss": -0.2933144826647762,
+      "reward": 1.71875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 1.0625,
+      "step": 2053
+    },
+    {
+      "epoch": 1.0970666666666666,
+      "grad_norm": 0.25678548216819763,
+      "learning_rate": 3e-06,
+      "loss": 0.0916,
+      "policy/loss": 0.01958085701466672,
+      "step": 2054
+    },
+    {
+      "epoch": 1.0976,
+      "grad_norm": 0.3066446781158447,
+      "learning_rate": 3e-06,
+      "loss": -0.004,
+      "policy/loss": -0.29615109756360525,
+      "step": 2055
+    },
+    {
+      "epoch": 1.0981333333333334,
+      "grad_norm": 0.24675941467285156,
+      "learning_rate": 3e-06,
+      "loss": 0.0913,
+      "policy/loss": 0.018188499993524232,
+      "step": 2056
+    },
+    {
+      "completion_length": 204.25000762939453,
+      "epoch": 1.0986666666666667,
+      "grad_norm": 0.2683275043964386,
+      "learning_rate": 3e-06,
+      "loss": -0.0203,
+      "policy/loss": -0.21416260674595833,
+      "reward": 1.7812501192092896,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 1.1250000298023224,
+      "step": 2057
+    },
+    {
+      "epoch": 1.0992,
+      "grad_norm": 0.3256928324699402,
+      "learning_rate": 3e-06,
+      "loss": -0.0367,
+      "policy/loss": -0.49959897994995117,
+      "step": 2058
+    },
+    {
+      "epoch": 1.0997333333333332,
+      "grad_norm": 0.2609912157058716,
+      "learning_rate": 3e-06,
+      "loss": -0.0215,
+      "policy/loss": -0.21786609664559364,
+      "step": 2059
+    },
+    {
+      "epoch": 1.1002666666666667,
+      "grad_norm": 0.32272613048553467,
+      "learning_rate": 3e-06,
+      "loss": -0.0386,
+      "policy/loss": -0.4989207237958908,
+      "step": 2060
+    },
+    {
+      "completion_length": 223.14583587646484,
+      "epoch": 1.1008,
+      "grad_norm": 0.28954821825027466,
+      "learning_rate": 3e-06,
+      "loss": -0.1533,
+      "policy/loss": -0.16560492205580601,
+      "reward": 1.3177083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.6041666567325592,
+      "step": 2061
+    },
+    {
+      "epoch": 1.1013333333333333,
+      "grad_norm": 0.28217077255249023,
+      "learning_rate": 3e-06,
+      "loss": -0.1497,
+      "policy/loss": -0.28352924288159187,
+      "step": 2062
+    },
+    {
+      "epoch": 1.1018666666666665,
+      "grad_norm": 0.2966883182525635,
+      "learning_rate": 3e-06,
+      "loss": -0.1545,
+      "policy/loss": -0.16499236517435278,
+      "step": 2063
+    },
+    {
+      "epoch": 1.1024,
+      "grad_norm": 0.2754587233066559,
+      "learning_rate": 3e-06,
+      "loss": -0.1508,
+      "policy/loss": -0.2851219386300272,
+      "step": 2064
+    },
+    {
+      "completion_length": 222.09375762939453,
+      "epoch": 1.1029333333333333,
+      "grad_norm": 0.2900453507900238,
+      "learning_rate": 3e-06,
+      "loss": -0.0248,
+      "policy/loss": 0.3294939696788788,
+      "reward": 1.5000000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 2065
+    },
+    {
+      "epoch": 1.1034666666666666,
+      "grad_norm": 0.24322372674942017,
+      "learning_rate": 3e-06,
+      "loss": -0.2716,
+      "policy/loss": -0.09192700684070587,
+      "step": 2066
+    },
+    {
+      "epoch": 1.104,
+      "grad_norm": 0.26508936285972595,
+      "learning_rate": 3e-06,
+      "loss": -0.0269,
+      "policy/loss": 0.32585084438323975,
+      "step": 2067
+    },
+    {
+      "epoch": 1.1045333333333334,
+      "grad_norm": 0.2793661952018738,
+      "learning_rate": 3e-06,
+      "loss": -0.2725,
+      "policy/loss": -0.09168067947030067,
+      "step": 2068
+    },
+    {
+      "completion_length": 218.80209350585938,
+      "epoch": 1.1050666666666666,
+      "grad_norm": 0.2565886378288269,
+      "learning_rate": 3e-06,
+      "loss": -0.1591,
+      "policy/loss": 1.0774848213657151e-07,
+      "reward": 1.8072917461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 1.0625000298023224,
+      "step": 2069
+    },
+    {
+      "epoch": 1.1056,
+      "grad_norm": 0.2622310221195221,
+      "learning_rate": 3e-06,
+      "loss": 0.0327,
+      "policy/loss": 1.0966229169184771e-07,
+      "step": 2070
+    },
+    {
+      "epoch": 1.1061333333333334,
+      "grad_norm": 0.24093176424503326,
+      "learning_rate": 3e-06,
+      "loss": -0.1601,
+      "policy/loss": 1.0786767035142475e-07,
+      "step": 2071
+    },
+    {
+      "epoch": 1.1066666666666667,
+      "grad_norm": 0.2650788426399231,
+      "learning_rate": 3e-06,
+      "loss": 0.0308,
+      "policy/loss": 1.0979011477729728e-07,
+      "step": 2072
+    },
+    {
+      "completion_length": 209.90625762939453,
+      "epoch": 1.1072,
+      "grad_norm": 0.4958108067512512,
+      "learning_rate": 3e-06,
+      "loss": -0.092,
+      "policy/loss": 0.09842174874886922,
+      "reward": 1.7864583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 1.1041666865348816,
+      "step": 2073
+    },
+    {
+      "epoch": 1.1077333333333332,
+      "grad_norm": 0.3223077654838562,
+      "learning_rate": 3e-06,
+      "loss": -0.2008,
+      "policy/loss": -0.08219071612971618,
+      "step": 2074
+    },
+    {
+      "epoch": 1.1082666666666667,
+      "grad_norm": 0.4714907705783844,
+      "learning_rate": 3e-06,
+      "loss": -0.0938,
+      "policy/loss": 0.08804689705358015,
+      "step": 2075
+    },
+    {
+      "epoch": 1.1088,
+      "grad_norm": 0.3025578558444977,
+      "learning_rate": 3e-06,
+      "loss": -0.2015,
+      "policy/loss": -0.0828183011466308,
+      "step": 2076
+    },
+    {
+      "completion_length": 206.80209350585938,
+      "epoch": 1.1093333333333333,
+      "grad_norm": 0.36459386348724365,
+      "learning_rate": 3e-06,
+      "loss": -0.1055,
+      "policy/loss": -0.15544317286298792,
+      "reward": 1.5260416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 2077
+    },
+    {
+      "epoch": 1.1098666666666666,
+      "grad_norm": 0.3985949158668518,
+      "learning_rate": 3e-06,
+      "loss": 0.0188,
+      "policy/loss": 0.38656446711516423,
+      "step": 2078
+    },
+    {
+      "epoch": 1.1104,
+      "grad_norm": 0.3719296455383301,
+      "learning_rate": 3e-06,
+      "loss": -0.108,
+      "policy/loss": -0.16377050791243875,
+      "step": 2079
+    },
+    {
+      "epoch": 1.1109333333333333,
+      "grad_norm": 0.3013918399810791,
+      "learning_rate": 3e-06,
+      "loss": 0.0177,
+      "policy/loss": 0.3809636272688479,
+      "step": 2080
+    },
+    {
+      "completion_length": 225.5104217529297,
+      "epoch": 1.1114666666666666,
+      "grad_norm": 0.2741020917892456,
+      "learning_rate": 3e-06,
+      "loss": -0.1092,
+      "policy/loss": -0.3514417037367821,
+      "reward": 1.3645833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 2081
+    },
+    {
+      "epoch": 1.112,
+      "grad_norm": 0.2696226239204407,
+      "learning_rate": 3e-06,
+      "loss": 0.0528,
+      "policy/loss": -0.24820588156580925,
+      "step": 2082
+    },
+    {
+      "epoch": 1.1125333333333334,
+      "grad_norm": 0.3031225800514221,
+      "learning_rate": 3e-06,
+      "loss": -0.1104,
+      "policy/loss": -0.35286543518304825,
+      "step": 2083
+    },
+    {
+      "epoch": 1.1130666666666666,
+      "grad_norm": 0.27100077271461487,
+      "learning_rate": 3e-06,
+      "loss": 0.0513,
+      "policy/loss": -0.253774493932724,
+      "step": 2084
+    },
+    {
+      "completion_length": 224.17708587646484,
+      "epoch": 1.1136,
+      "grad_norm": 1.073825716972351,
+      "learning_rate": 3e-06,
+      "loss": -0.1899,
+      "policy/loss": 0.11781414381430366,
+      "reward": 1.4166667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333134651184,
+      "rewards/correctness_reward_func_math": 0.7083333134651184,
+      "step": 2085
+    },
+    {
+      "epoch": 1.1141333333333334,
+      "grad_norm": 0.30280768871307373,
+      "learning_rate": 3e-06,
+      "loss": -0.036,
+      "policy/loss": 0.004992043792903189,
+      "step": 2086
+    },
+    {
+      "epoch": 1.1146666666666667,
+      "grad_norm": 0.2754865288734436,
+      "learning_rate": 3e-06,
+      "loss": -0.1889,
+      "policy/loss": 0.1181237899840184,
+      "step": 2087
+    },
+    {
+      "epoch": 1.1152,
+      "grad_norm": 0.282929927110672,
+      "learning_rate": 3e-06,
+      "loss": -0.038,
+      "policy/loss": 0.0045017737960755255,
+      "step": 2088
+    },
+    {
+      "completion_length": 191.15625762939453,
+      "epoch": 1.1157333333333332,
+      "grad_norm": 0.27780619263648987,
+      "learning_rate": 3e-06,
+      "loss": 0.0076,
+      "policy/loss": 6.933900564831674e-08,
+      "reward": 1.7187501192092896,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 1.0000000298023224,
+      "step": 2089
+    },
+    {
+      "epoch": 1.1162666666666667,
+      "grad_norm": 0.270494282245636,
+      "learning_rate": 3e-06,
+      "loss": 0.0036,
+      "policy/loss": 6.76983820113719e-08,
+      "step": 2090
+    },
+    {
+      "epoch": 1.1168,
+      "grad_norm": 0.246896430850029,
+      "learning_rate": 3e-06,
+      "loss": 0.0059,
+      "policy/loss": 6.936248908573361e-08,
+      "step": 2091
+    },
+    {
+      "epoch": 1.1173333333333333,
+      "grad_norm": 0.3143901824951172,
+      "learning_rate": 3e-06,
+      "loss": 0.0031,
+      "policy/loss": 6.79423948213298e-08,
+      "step": 2092
+    },
+    {
+      "completion_length": 210.0416717529297,
+      "epoch": 1.1178666666666666,
+      "grad_norm": 0.256059855222702,
+      "learning_rate": 3e-06,
+      "loss": 0.1456,
+      "policy/loss": 5.476166009543704e-08,
+      "reward": 1.6875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.9375,
+      "step": 2093
+    },
+    {
+      "epoch": 1.1184,
+      "grad_norm": 0.3007746636867523,
+      "learning_rate": 3e-06,
+      "loss": 0.067,
+      "policy/loss": 8.778479809734563e-08,
+      "step": 2094
+    },
+    {
+      "epoch": 1.1189333333333333,
+      "grad_norm": 0.24716034531593323,
+      "learning_rate": 3e-06,
+      "loss": 0.1457,
+      "policy/loss": 5.475522080189421e-08,
+      "step": 2095
+    },
+    {
+      "epoch": 1.1194666666666666,
+      "grad_norm": 0.28995537757873535,
+      "learning_rate": 3e-06,
+      "loss": 0.0655,
+      "policy/loss": 8.801776729683297e-08,
+      "step": 2096
+    },
+    {
+      "completion_length": 231.52083587646484,
+      "epoch": 1.12,
+      "grad_norm": 0.3488827347755432,
+      "learning_rate": 3e-06,
+      "loss": -0.0088,
+      "policy/loss": 6.376683359121671e-08,
+      "reward": 1.1770833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.59375,
+      "rewards/correctness_reward_func_math": 0.5833333432674408,
+      "step": 2097
+    },
+    {
+      "epoch": 1.1205333333333334,
+      "grad_norm": 0.2037874013185501,
+      "learning_rate": 3e-06,
+      "loss": 0.2899,
+      "policy/loss": 7.076753405499403e-08,
+      "step": 2098
+    },
+    {
+      "epoch": 1.1210666666666667,
+      "grad_norm": 0.3882960081100464,
+      "learning_rate": 3e-06,
+      "loss": -0.0108,
+      "policy/loss": 6.358509807569135e-08,
+      "step": 2099
+    },
+    {
+      "epoch": 1.1216,
+      "grad_norm": 0.22829246520996094,
+      "learning_rate": 3e-06,
+      "loss": 0.289,
+      "policy/loss": 7.069392893299664e-08,
+      "step": 2100
+    },
+    {
+      "completion_length": 207.6666717529297,
+      "epoch": 1.1221333333333334,
+      "grad_norm": 0.3003827631473541,
+      "learning_rate": 3e-06,
+      "loss": 0.1789,
+      "policy/loss": 0.9543245069720978,
+      "reward": 2.0,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 1.3125,
+      "step": 2101
+    },
+    {
+      "epoch": 1.1226666666666667,
+      "grad_norm": 0.32321903109550476,
+      "learning_rate": 3e-06,
+      "loss": 0.2844,
+      "policy/loss": 0.84000286473416,
+      "step": 2102
+    },
+    {
+      "epoch": 1.1232,
+      "grad_norm": 0.2840973138809204,
+      "learning_rate": 3e-06,
+      "loss": 0.1774,
+      "policy/loss": 0.9495488636124172,
+      "step": 2103
+    },
+    {
+      "epoch": 1.1237333333333333,
+      "grad_norm": 0.2937942445278168,
+      "learning_rate": 3e-06,
+      "loss": 0.2822,
+      "policy/loss": 0.837272614332651,
+      "step": 2104
+    },
+    {
+      "completion_length": 217.6041717529297,
+      "epoch": 1.1242666666666667,
+      "grad_norm": 0.26461511850357056,
+      "learning_rate": 3e-06,
+      "loss": 0.0905,
+      "policy/loss": 0.048815403578595706,
+      "reward": 1.4687500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 2105
+    },
+    {
+      "epoch": 1.1248,
+      "grad_norm": 0.39216718077659607,
+      "learning_rate": 3e-06,
+      "loss": 0.1106,
+      "policy/loss": 0.41202663124990835,
+      "step": 2106
+    },
+    {
+      "epoch": 1.1253333333333333,
+      "grad_norm": 0.2438002973794937,
+      "learning_rate": 3e-06,
+      "loss": 0.0899,
+      "policy/loss": 0.04842950074078267,
+      "step": 2107
+    },
+    {
+      "epoch": 1.1258666666666666,
+      "grad_norm": 0.4779147803783417,
+      "learning_rate": 3e-06,
+      "loss": 0.1106,
+      "policy/loss": 0.4151921743755107,
+      "step": 2108
+    },
+    {
+      "completion_length": 216.9375,
+      "epoch": 1.1264,
+      "grad_norm": 0.2482011914253235,
+      "learning_rate": 3e-06,
+      "loss": 0.0347,
+      "policy/loss": -0.11998351183153844,
+      "reward": 1.7135417461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 1.0624999701976776,
+      "step": 2109
+    },
+    {
+      "epoch": 1.1269333333333333,
+      "grad_norm": 0.2769639194011688,
+      "learning_rate": 3e-06,
+      "loss": 0.0799,
+      "policy/loss": -0.08914470126413931,
+      "step": 2110
+    },
+    {
+      "epoch": 1.1274666666666666,
+      "grad_norm": 0.2858078181743622,
+      "learning_rate": 3e-06,
+      "loss": 0.0342,
+      "policy/loss": -0.11920286990475049,
+      "step": 2111
+    },
+    {
+      "epoch": 1.1280000000000001,
+      "grad_norm": 0.2544654309749603,
+      "learning_rate": 3e-06,
+      "loss": 0.0782,
+      "policy/loss": -0.09096797752285113,
+      "step": 2112
+    },
+    {
+      "completion_length": 208.6666717529297,
+      "epoch": 1.1285333333333334,
+      "grad_norm": 0.35802170634269714,
+      "learning_rate": 3e-06,
+      "loss": 0.0566,
+      "policy/loss": 0.09846971277147532,
+      "reward": 1.4583333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.7083333283662796,
+      "step": 2113
+    },
+    {
+      "epoch": 1.1290666666666667,
+      "grad_norm": 0.26756325364112854,
+      "learning_rate": 3e-06,
+      "loss": 0.0595,
+      "policy/loss": 0.248337734490633,
+      "step": 2114
+    },
+    {
+      "epoch": 1.1296,
+      "grad_norm": 0.2940458059310913,
+      "learning_rate": 3e-06,
+      "loss": 0.0554,
+      "policy/loss": 0.09701946377754211,
+      "step": 2115
+    },
+    {
+      "epoch": 1.1301333333333332,
+      "grad_norm": 0.26200148463249207,
+      "learning_rate": 3e-06,
+      "loss": 0.0584,
+      "policy/loss": 0.24549296870827675,
+      "step": 2116
+    },
+    {
+      "completion_length": 220.7291717529297,
+      "epoch": 1.1306666666666667,
+      "grad_norm": 0.2828771770000458,
+      "learning_rate": 3e-06,
+      "loss": 0.131,
+      "policy/loss": 1.1349527895276879e-07,
+      "reward": 1.2968750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.5833333283662796,
+      "step": 2117
+    },
+    {
+      "epoch": 1.1312,
+      "grad_norm": 0.3426845073699951,
+      "learning_rate": 3e-06,
+      "loss": 0.0952,
+      "policy/loss": 1.0302755626412363e-07,
+      "step": 2118
+    },
+    {
+      "epoch": 1.1317333333333333,
+      "grad_norm": 0.26385045051574707,
+      "learning_rate": 3e-06,
+      "loss": 0.1298,
+      "policy/loss": 1.1352723916502327e-07,
+      "step": 2119
+    },
+    {
+      "epoch": 1.1322666666666668,
+      "grad_norm": 0.33291736245155334,
+      "learning_rate": 3e-06,
+      "loss": 0.0927,
+      "policy/loss": 1.0324387034188476e-07,
+      "step": 2120
+    },
+    {
+      "completion_length": 205.80209350585938,
+      "epoch": 1.1328,
+      "grad_norm": 0.3040824234485626,
+      "learning_rate": 3e-06,
+      "loss": -0.1478,
+      "policy/loss": 4.033819323012722e-08,
+      "reward": 1.3697916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.703125,
+      "rewards/correctness_reward_func_math": 0.6666666641831398,
+      "step": 2121
+    },
+    {
+      "epoch": 1.1333333333333333,
+      "grad_norm": 0.22882723808288574,
+      "learning_rate": 3e-06,
+      "loss": -0.1054,
+      "policy/loss": 5.117414048072533e-08,
+      "step": 2122
+    },
+    {
+      "epoch": 1.1338666666666666,
+      "grad_norm": 0.27349281311035156,
+      "learning_rate": 3e-06,
+      "loss": -0.1497,
+      "policy/loss": 4.0335123685508734e-08,
+      "step": 2123
+    },
+    {
+      "epoch": 1.1344,
+      "grad_norm": 0.24806907773017883,
+      "learning_rate": 3e-06,
+      "loss": -0.107,
+      "policy/loss": 5.1116071375645333e-08,
+      "step": 2124
+    },
+    {
+      "completion_length": 210.11458587646484,
+      "epoch": 1.1349333333333333,
+      "grad_norm": 0.25938138365745544,
+      "learning_rate": 3e-06,
+      "loss": 0.0395,
+      "policy/loss": 1.0942817496584212e-07,
+      "reward": 1.2916666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.5416666567325592,
+      "step": 2125
+    },
+    {
+      "epoch": 1.1354666666666666,
+      "grad_norm": 0.2947342097759247,
+      "learning_rate": 3e-06,
+      "loss": 0.0146,
+      "policy/loss": 8.429742237581195e-08,
+      "step": 2126
+    },
+    {
+      "epoch": 1.1360000000000001,
+      "grad_norm": 0.2695062458515167,
+      "learning_rate": 3e-06,
+      "loss": 0.0382,
+      "policy/loss": 1.0938214156652748e-07,
+      "step": 2127
+    },
+    {
+      "epoch": 1.1365333333333334,
+      "grad_norm": 0.2886114716529846,
+      "learning_rate": 3e-06,
+      "loss": 0.013,
+      "policy/loss": 8.444007626451366e-08,
+      "step": 2128
+    },
+    {
+      "completion_length": 222.59375,
+      "epoch": 1.1370666666666667,
+      "grad_norm": 0.29575273394584656,
+      "learning_rate": 3e-06,
+      "loss": 0.1255,
+      "policy/loss": 0.10064301921139673,
+      "reward": 1.8229166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 1.1041666865348816,
+      "step": 2129
+    },
+    {
+      "epoch": 1.1376,
+      "grad_norm": 0.2551318109035492,
+      "learning_rate": 3e-06,
+      "loss": 0.0151,
+      "policy/loss": 0.2154703940986451,
+      "step": 2130
+    },
+    {
+      "epoch": 1.1381333333333332,
+      "grad_norm": 0.2919102907180786,
+      "learning_rate": 3e-06,
+      "loss": 0.1246,
+      "policy/loss": 0.1007600976994425,
+      "step": 2131
+    },
+    {
+      "epoch": 1.1386666666666667,
+      "grad_norm": 0.24938276410102844,
+      "learning_rate": 3e-06,
+      "loss": 0.0141,
+      "policy/loss": 0.21555746132011677,
+      "step": 2132
+    },
+    {
+      "completion_length": 203.55208587646484,
+      "epoch": 1.1392,
+      "grad_norm": 0.26316601037979126,
+      "learning_rate": 3e-06,
+      "loss": -0.0249,
+      "policy/loss": 6.131692487088003e-08,
+      "reward": 1.6302083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.9166666567325592,
+      "step": 2133
+    },
+    {
+      "epoch": 1.1397333333333333,
+      "grad_norm": 0.20092622935771942,
+      "learning_rate": 3e-06,
+      "loss": -0.0237,
+      "policy/loss": 5.90343240958191e-08,
+      "step": 2134
+    },
+    {
+      "epoch": 1.1402666666666668,
+      "grad_norm": 0.2766530513763428,
+      "learning_rate": 3e-06,
+      "loss": -0.026,
+      "policy/loss": 6.132119523272195e-08,
+      "step": 2135
+    },
+    {
+      "epoch": 1.1408,
+      "grad_norm": 0.21694722771644592,
+      "learning_rate": 3e-06,
+      "loss": -0.0244,
+      "policy/loss": 5.9063165025463604e-08,
+      "step": 2136
+    },
+    {
+      "completion_length": 220.09375762939453,
+      "epoch": 1.1413333333333333,
+      "grad_norm": 0.2851365804672241,
+      "learning_rate": 3e-06,
+      "loss": -0.0812,
+      "policy/loss": 9.388638844143316e-08,
+      "reward": 1.182291716337204,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5989583432674408,
+      "rewards/correctness_reward_func_math": 0.5833333432674408,
+      "step": 2137
+    },
+    {
+      "epoch": 1.1418666666666666,
+      "grad_norm": 0.28326067328453064,
+      "learning_rate": 3e-06,
+      "loss": 0.0281,
+      "policy/loss": 9.789337696020084e-08,
+      "step": 2138
+    },
+    {
+      "epoch": 1.1424,
+      "grad_norm": 0.30593112111091614,
+      "learning_rate": 3e-06,
+      "loss": -0.084,
+      "policy/loss": 9.399580491731285e-08,
+      "step": 2139
+    },
+    {
+      "epoch": 1.1429333333333334,
+      "grad_norm": 0.28731289505958557,
+      "learning_rate": 3e-06,
+      "loss": 0.027,
+      "policy/loss": 9.774392495387474e-08,
+      "step": 2140
+    },
+    {
+      "completion_length": 208.8541717529297,
+      "epoch": 1.1434666666666666,
+      "grad_norm": 0.2690492868423462,
+      "learning_rate": 3e-06,
+      "loss": 0.1078,
+      "policy/loss": 1.0121519267158874e-07,
+      "reward": 1.5625000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6666666567325592,
+      "rewards/correctness_reward_func_math": 0.8958333432674408,
+      "step": 2141
+    },
+    {
+      "epoch": 1.144,
+      "grad_norm": 0.371646523475647,
+      "learning_rate": 3e-06,
+      "loss": -0.0012,
+      "policy/loss": 1.0736307842762471e-07,
+      "step": 2142
+    },
+    {
+      "epoch": 1.1445333333333334,
+      "grad_norm": 0.27415311336517334,
+      "learning_rate": 3e-06,
+      "loss": 0.1077,
+      "policy/loss": 1.0118903048805805e-07,
+      "step": 2143
+    },
+    {
+      "epoch": 1.1450666666666667,
+      "grad_norm": 0.3859700858592987,
+      "learning_rate": 3e-06,
+      "loss": -0.002,
+      "policy/loss": 1.071873008129387e-07,
+      "step": 2144
+    },
+    {
+      "completion_length": 212.77083587646484,
+      "epoch": 1.1456,
+      "grad_norm": 0.2365172952413559,
+      "learning_rate": 3e-06,
+      "loss": -0.1712,
+      "policy/loss": 2.0202072548158867e-07,
+      "reward": 1.4531250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.7708333134651184,
+      "step": 2145
+    },
+    {
+      "epoch": 1.1461333333333332,
+      "grad_norm": 0.2996065616607666,
+      "learning_rate": 3e-06,
+      "loss": 0.077,
+      "policy/loss": 1.66887204500199e-07,
+      "step": 2146
+    },
+    {
+      "epoch": 1.1466666666666667,
+      "grad_norm": 0.25084125995635986,
+      "learning_rate": 3e-06,
+      "loss": -0.1728,
+      "policy/loss": 2.0210330120562503e-07,
+      "step": 2147
+    },
+    {
+      "epoch": 1.1472,
+      "grad_norm": 0.3143674433231354,
+      "learning_rate": 3e-06,
+      "loss": 0.0733,
+      "policy/loss": 1.664643214382977e-07,
+      "step": 2148
+    },
+    {
+      "completion_length": 213.1041717529297,
+      "epoch": 1.1477333333333333,
+      "grad_norm": 0.20691831409931183,
+      "learning_rate": 3e-06,
+      "loss": -0.1239,
+      "policy/loss": -0.5706852078437805,
+      "reward": 1.2708333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.5833333432674408,
+      "step": 2149
+    },
+    {
+      "epoch": 1.1482666666666668,
+      "grad_norm": 0.17929789423942566,
+      "learning_rate": 3e-06,
+      "loss": -0.0786,
+      "policy/loss": -0.6091777682304382,
+      "step": 2150
+    },
+    {
+      "epoch": 1.1488,
+      "grad_norm": 0.2128952592611313,
+      "learning_rate": 3e-06,
+      "loss": -0.1245,
+      "policy/loss": -0.5743197202682495,
+      "step": 2151
+    },
+    {
+      "epoch": 1.1493333333333333,
+      "grad_norm": 0.18149849772453308,
+      "learning_rate": 3e-06,
+      "loss": -0.0793,
+      "policy/loss": -0.6134032011032104,
+      "step": 2152
+    },
+    {
+      "completion_length": 209.5,
+      "epoch": 1.1498666666666666,
+      "grad_norm": 0.2015146017074585,
+      "learning_rate": 3e-06,
+      "loss": -0.0393,
+      "policy/loss": 4.7329631769343905e-08,
+      "reward": 1.3072916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5989583432674408,
+      "rewards/correctness_reward_func_math": 0.7083333432674408,
+      "step": 2153
+    },
+    {
+      "epoch": 1.1504,
+      "grad_norm": 0.3846926987171173,
+      "learning_rate": 3e-06,
+      "loss": -0.0144,
+      "policy/loss": 4.9686864400655395e-08,
+      "step": 2154
+    },
+    {
+      "epoch": 1.1509333333333334,
+      "grad_norm": 0.19365699589252472,
+      "learning_rate": 3e-06,
+      "loss": -0.0398,
+      "policy/loss": 4.740721237794787e-08,
+      "step": 2155
+    },
+    {
+      "epoch": 1.1514666666666666,
+      "grad_norm": 0.37478598952293396,
+      "learning_rate": 3e-06,
+      "loss": -0.0142,
+      "policy/loss": 4.985579593608236e-08,
+      "step": 2156
+    },
+    {
+      "completion_length": 197.28125762939453,
+      "epoch": 1.152,
+      "grad_norm": 0.3453645408153534,
+      "learning_rate": 3e-06,
+      "loss": 0.0947,
+      "policy/loss": 1.3749662741702195e-07,
+      "reward": 1.3906250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.7083333134651184,
+      "step": 2157
+    },
+    {
+      "epoch": 1.1525333333333334,
+      "grad_norm": 0.2725186049938202,
+      "learning_rate": 3e-06,
+      "loss": 0.139,
+      "policy/loss": 1.0309009823572524e-07,
+      "step": 2158
+    },
+    {
+      "epoch": 1.1530666666666667,
+      "grad_norm": 0.35324275493621826,
+      "learning_rate": 3e-06,
+      "loss": 0.0934,
+      "policy/loss": 1.3736794457486212e-07,
+      "step": 2159
+    },
+    {
+      "epoch": 1.1536,
+      "grad_norm": 0.2696947157382965,
+      "learning_rate": 3e-06,
+      "loss": 0.1377,
+      "policy/loss": 1.030849929861688e-07,
+      "step": 2160
+    },
+    {
+      "completion_length": 201.1979217529297,
+      "epoch": 1.1541333333333332,
+      "grad_norm": 0.24098215997219086,
+      "learning_rate": 3e-06,
+      "loss": -0.0607,
+      "policy/loss": 0.21561912600985167,
+      "reward": 1.7708333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 1.0833333134651184,
+      "step": 2161
+    },
+    {
+      "epoch": 1.1546666666666667,
+      "grad_norm": 0.24946604669094086,
+      "learning_rate": 3e-06,
+      "loss": 0.0963,
+      "policy/loss": -0.06949677785596364,
+      "step": 2162
+    },
+    {
+      "epoch": 1.1552,
+      "grad_norm": 0.2496003657579422,
+      "learning_rate": 3e-06,
+      "loss": -0.0617,
+      "policy/loss": 0.21069719803654152,
+      "step": 2163
+    },
+    {
+      "epoch": 1.1557333333333333,
+      "grad_norm": 0.25356683135032654,
+      "learning_rate": 3e-06,
+      "loss": 0.0957,
+      "policy/loss": -0.07426073857240922,
+      "step": 2164
+    },
+    {
+      "completion_length": 222.4375,
+      "epoch": 1.1562666666666668,
+      "grad_norm": 0.28308960795402527,
+      "learning_rate": 3e-06,
+      "loss": 0.1193,
+      "policy/loss": 0.16423114440868858,
+      "reward": 1.0416666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.4166666567325592,
+      "step": 2165
+    },
+    {
+      "epoch": 1.1568,
+      "grad_norm": 0.24730776250362396,
+      "learning_rate": 3e-06,
+      "loss": 0.0758,
+      "policy/loss": 0.28024377497423103,
+      "step": 2166
+    },
+    {
+      "epoch": 1.1573333333333333,
+      "grad_norm": 0.2768745422363281,
+      "learning_rate": 3e-06,
+      "loss": 0.1185,
+      "policy/loss": 0.16429851253278116,
+      "step": 2167
+    },
+    {
+      "epoch": 1.1578666666666666,
+      "grad_norm": 0.2594543993473053,
+      "learning_rate": 3e-06,
+      "loss": 0.0757,
+      "policy/loss": 0.2788741496536673,
+      "step": 2168
+    },
+    {
+      "completion_length": 204.27083587646484,
+      "epoch": 1.1584,
+      "grad_norm": 0.21436721086502075,
+      "learning_rate": 3e-06,
+      "loss": -0.0965,
+      "policy/loss": 7.925437017775039e-08,
+      "reward": 1.6041666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.9166666865348816,
+      "step": 2169
+    },
+    {
+      "epoch": 1.1589333333333334,
+      "grad_norm": 0.250800222158432,
+      "learning_rate": 3e-06,
+      "loss": 0.0381,
+      "policy/loss": 8.885243829581668e-08,
+      "step": 2170
+    },
+    {
+      "epoch": 1.1594666666666666,
+      "grad_norm": 0.21417412161827087,
+      "learning_rate": 3e-06,
+      "loss": -0.0964,
+      "policy/loss": 7.939471657891772e-08,
+      "step": 2171
+    },
+    {
+      "epoch": 1.16,
+      "grad_norm": 0.249708354473114,
+      "learning_rate": 3e-06,
+      "loss": 0.0377,
+      "policy/loss": 8.852835264860914e-08,
+      "step": 2172
+    },
+    {
+      "completion_length": 206.11458587646484,
+      "epoch": 1.1605333333333334,
+      "grad_norm": 0.3252304494380951,
+      "learning_rate": 3e-06,
+      "loss": -0.0906,
+      "policy/loss": -0.4573926144204137,
+      "reward": 1.4114583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6197916567325592,
+      "rewards/correctness_reward_func_math": 0.7916666865348816,
+      "step": 2173
+    },
+    {
+      "epoch": 1.1610666666666667,
+      "grad_norm": 0.3182126581668854,
+      "learning_rate": 3e-06,
+      "loss": -0.0842,
+      "policy/loss": -0.1272141606533701,
+      "step": 2174
+    },
+    {
+      "epoch": 1.1616,
+      "grad_norm": 0.32582738995552063,
+      "learning_rate": 3e-06,
+      "loss": -0.0917,
+      "policy/loss": -0.4555893057241356,
+      "step": 2175
+    },
+    {
+      "epoch": 1.1621333333333332,
+      "grad_norm": 0.3206765055656433,
+      "learning_rate": 3e-06,
+      "loss": -0.0852,
+      "policy/loss": -0.12949937288898639,
+      "step": 2176
+    },
+    {
+      "completion_length": 204.55209350585938,
+      "epoch": 1.1626666666666667,
+      "grad_norm": 0.2652207016944885,
+      "learning_rate": 3e-06,
+      "loss": -0.1425,
+      "policy/loss": -0.0575173944234848,
+      "reward": 1.3333333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 2177
+    },
+    {
+      "epoch": 1.1632,
+      "grad_norm": 0.2704797089099884,
+      "learning_rate": 3e-06,
+      "loss": -0.1031,
+      "policy/loss": -0.6657322496175766,
+      "step": 2178
+    },
+    {
+      "epoch": 1.1637333333333333,
+      "grad_norm": 0.26126664876937866,
+      "learning_rate": 3e-06,
+      "loss": -0.1428,
+      "policy/loss": -0.05818690359592438,
+      "step": 2179
+    },
+    {
+      "epoch": 1.1642666666666668,
+      "grad_norm": 0.30351534485816956,
+      "learning_rate": 3e-06,
+      "loss": -0.105,
+      "policy/loss": -0.6703606396913528,
+      "step": 2180
+    },
+    {
+      "completion_length": 221.4791717529297,
+      "epoch": 1.1648,
+      "grad_norm": 0.468295156955719,
+      "learning_rate": 3e-06,
+      "loss": -0.2132,
+      "policy/loss": 1.0259593707928616e-07,
+      "reward": 1.3385416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6302083432674408,
+      "rewards/correctness_reward_func_math": 0.708333358168602,
+      "step": 2181
+    },
+    {
+      "epoch": 1.1653333333333333,
+      "grad_norm": 0.2964112460613251,
+      "learning_rate": 3e-06,
+      "loss": -0.1279,
+      "policy/loss": 9.092405761634836e-08,
+      "step": 2182
+    },
+    {
+      "epoch": 1.1658666666666666,
+      "grad_norm": 0.48511549830436707,
+      "learning_rate": 3e-06,
+      "loss": -0.2132,
+      "policy/loss": 1.0256703575350912e-07,
+      "step": 2183
+    },
+    {
+      "epoch": 1.1663999999999999,
+      "grad_norm": 0.33043748140335083,
+      "learning_rate": 3e-06,
+      "loss": -0.1284,
+      "policy/loss": 9.09531436832367e-08,
+      "step": 2184
+    },
+    {
+      "completion_length": 200.65625,
+      "epoch": 1.1669333333333334,
+      "grad_norm": 0.3044625520706177,
+      "learning_rate": 3e-06,
+      "loss": 0.0291,
+      "policy/loss": 1.5262688179973338e-07,
+      "reward": 1.8333333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 1.1458333134651184,
+      "step": 2185
+    },
+    {
+      "epoch": 1.1674666666666667,
+      "grad_norm": 0.3577664792537689,
+      "learning_rate": 3e-06,
+      "loss": -0.0315,
+      "policy/loss": 1.3914148766502876e-07,
+      "step": 2186
+    },
+    {
+      "epoch": 1.168,
+      "grad_norm": 0.286350816488266,
+      "learning_rate": 3e-06,
+      "loss": 0.027,
+      "policy/loss": 1.523995010188628e-07,
+      "step": 2187
+    },
+    {
+      "epoch": 1.1685333333333334,
+      "grad_norm": 0.35637375712394714,
+      "learning_rate": 3e-06,
+      "loss": -0.0318,
+      "policy/loss": 1.3936574561057569e-07,
+      "step": 2188
+    },
+    {
+      "completion_length": 209.95834350585938,
+      "epoch": 1.1690666666666667,
+      "grad_norm": 0.23191046714782715,
+      "learning_rate": 3e-06,
+      "loss": 0.1355,
+      "policy/loss": 0.003043672928441765,
+      "reward": 1.1614583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.4791666716337204,
+      "step": 2189
+    },
+    {
+      "epoch": 1.1696,
+      "grad_norm": 0.2404623031616211,
+      "learning_rate": 3e-06,
+      "loss": 0.0033,
+      "policy/loss": -0.0596372427505365,
+      "step": 2190
+    },
+    {
+      "epoch": 1.1701333333333332,
+      "grad_norm": 0.22965554893016815,
+      "learning_rate": 3e-06,
+      "loss": 0.1358,
+      "policy/loss": 0.00125171366586585,
+      "step": 2191
+    },
+    {
+      "epoch": 1.1706666666666667,
+      "grad_norm": 0.23225799202919006,
+      "learning_rate": 3e-06,
+      "loss": 0.0023,
+      "policy/loss": -0.060527553514280186,
+      "step": 2192
+    },
+    {
+      "completion_length": 205.5625,
+      "epoch": 1.1712,
+      "grad_norm": 0.2517227530479431,
+      "learning_rate": 3e-06,
+      "loss": -0.0791,
+      "policy/loss": -0.5971790701150894,
+      "reward": 1.6927083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 1.0416666865348816,
+      "step": 2193
+    },
+    {
+      "epoch": 1.1717333333333333,
+      "grad_norm": 0.2848559021949768,
+      "learning_rate": 3e-06,
+      "loss": -0.1095,
+      "policy/loss": 0.11885736882686615,
+      "step": 2194
+    },
+    {
+      "epoch": 1.1722666666666668,
+      "grad_norm": 0.2377357929944992,
+      "learning_rate": 3e-06,
+      "loss": -0.081,
+      "policy/loss": -0.6057026088237762,
+      "step": 2195
+    },
+    {
+      "epoch": 1.1728,
+      "grad_norm": 0.26871412992477417,
+      "learning_rate": 3e-06,
+      "loss": -0.1113,
+      "policy/loss": 0.11069972813129425,
+      "step": 2196
+    },
+    {
+      "completion_length": 202.4375,
+      "epoch": 1.1733333333333333,
+      "grad_norm": 0.32493728399276733,
+      "learning_rate": 3e-06,
+      "loss": 0.2095,
+      "policy/loss": 0.3376360833644867,
+      "reward": 1.8333333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333134651184,
+      "rewards/correctness_reward_func_math": 1.125,
+      "step": 2197
+    },
+    {
+      "epoch": 1.1738666666666666,
+      "grad_norm": 0.297908216714859,
+      "learning_rate": 3e-06,
+      "loss": 0.138,
+      "policy/loss": -0.07613147050142288,
+      "step": 2198
+    },
+    {
+      "epoch": 1.1743999999999999,
+      "grad_norm": 0.327981173992157,
+      "learning_rate": 3e-06,
+      "loss": 0.2087,
+      "policy/loss": 0.33573613315820694,
+      "step": 2199
+    },
+    {
+      "epoch": 1.1749333333333334,
+      "grad_norm": 0.2721838057041168,
+      "learning_rate": 3e-06,
+      "loss": 0.1374,
+      "policy/loss": -0.07977519556879997,
+      "step": 2200
+    },
+    {
+      "completion_length": 202.08334350585938,
+      "epoch": 1.1754666666666667,
+      "grad_norm": 0.36845308542251587,
+      "learning_rate": 3e-06,
+      "loss": -0.0543,
+      "policy/loss": -0.10808652931636686,
+      "reward": 1.7291666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.9791666567325592,
+      "step": 2201
+    },
+    {
+      "epoch": 1.176,
+      "grad_norm": 0.2943701446056366,
+      "learning_rate": 3e-06,
+      "loss": 0.2346,
+      "policy/loss": 0.1472814239817879,
+      "step": 2202
+    },
+    {
+      "epoch": 1.1765333333333334,
+      "grad_norm": 0.3324855864048004,
+      "learning_rate": 3e-06,
+      "loss": -0.0552,
+      "policy/loss": -0.11073375022691678,
+      "step": 2203
+    },
+    {
+      "epoch": 1.1770666666666667,
+      "grad_norm": 0.3095892071723938,
+      "learning_rate": 3e-06,
+      "loss": 0.2335,
+      "policy/loss": 0.15169045420710603,
+      "step": 2204
+    },
+    {
+      "completion_length": 214.2604217529297,
+      "epoch": 1.1776,
+      "grad_norm": 0.27346527576446533,
+      "learning_rate": 3e-06,
+      "loss": -0.006,
+      "policy/loss": 9.657230393145255e-08,
+      "reward": 1.1510416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.5,
+      "step": 2205
+    },
+    {
+      "epoch": 1.1781333333333333,
+      "grad_norm": 0.27948009967803955,
+      "learning_rate": 3e-06,
+      "loss": -0.0392,
+      "policy/loss": 9.900221087377759e-08,
+      "step": 2206
+    },
+    {
+      "epoch": 1.1786666666666668,
+      "grad_norm": 0.3755108714103699,
+      "learning_rate": 3e-06,
+      "loss": -0.007,
+      "policy/loss": 9.646747756164586e-08,
+      "step": 2207
+    },
+    {
+      "epoch": 1.1792,
+      "grad_norm": 0.277955561876297,
+      "learning_rate": 3e-06,
+      "loss": -0.0412,
+      "policy/loss": 9.901037500981147e-08,
+      "step": 2208
+    },
+    {
+      "completion_length": 200.06250762939453,
+      "epoch": 1.1797333333333333,
+      "grad_norm": 0.2673877477645874,
+      "learning_rate": 3e-06,
+      "loss": -0.0343,
+      "policy/loss": 0.255224464402124,
+      "reward": 1.8489583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 1.1041666269302368,
+      "step": 2209
+    },
+    {
+      "epoch": 1.1802666666666666,
+      "grad_norm": 0.28108328580856323,
+      "learning_rate": 3e-06,
+      "loss": 0.0179,
+      "policy/loss": 0.19824441134025506,
+      "step": 2210
+    },
+    {
+      "epoch": 1.1808,
+      "grad_norm": 0.25668200850486755,
+      "learning_rate": 3e-06,
+      "loss": -0.0354,
+      "policy/loss": 0.2555572073579153,
+      "step": 2211
+    },
+    {
+      "epoch": 1.1813333333333333,
+      "grad_norm": 0.2706082761287689,
+      "learning_rate": 3e-06,
+      "loss": 0.0167,
+      "policy/loss": 0.19720434006520016,
+      "step": 2212
+    },
+    {
+      "completion_length": 215.40625,
+      "epoch": 1.1818666666666666,
+      "grad_norm": 0.2906379699707031,
+      "learning_rate": 3e-06,
+      "loss": 0.0514,
+      "policy/loss": 0.08838319033384323,
+      "reward": 1.4583333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.7083333134651184,
+      "step": 2213
+    },
+    {
+      "epoch": 1.1824,
+      "grad_norm": 0.44758686423301697,
+      "learning_rate": 3e-06,
+      "loss": 0.0654,
+      "policy/loss": 0.002211064100265503,
+      "step": 2214
+    },
+    {
+      "epoch": 1.1829333333333334,
+      "grad_norm": 0.2710990607738495,
+      "learning_rate": 3e-06,
+      "loss": 0.0506,
+      "policy/loss": 0.0876937210559845,
+      "step": 2215
+    },
+    {
+      "epoch": 1.1834666666666667,
+      "grad_norm": 0.4384620785713196,
+      "learning_rate": 3e-06,
+      "loss": 0.063,
+      "policy/loss": -0.0020395219326019287,
+      "step": 2216
+    },
+    {
+      "completion_length": 212.3541717529297,
+      "epoch": 1.184,
+      "grad_norm": 0.22848692536354065,
+      "learning_rate": 3e-06,
+      "loss": -0.0339,
+      "policy/loss": 0.29609096730348483,
+      "reward": 1.4583333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7708333432674408,
+      "step": 2217
+    },
+    {
+      "epoch": 1.1845333333333334,
+      "grad_norm": 0.2880021035671234,
+      "learning_rate": 3e-06,
+      "loss": 0.0741,
+      "policy/loss": 0.6637853255644828,
+      "step": 2218
+    },
+    {
+      "epoch": 1.1850666666666667,
+      "grad_norm": 0.23301491141319275,
+      "learning_rate": 3e-06,
+      "loss": -0.0346,
+      "policy/loss": 0.29610190479574783,
+      "step": 2219
+    },
+    {
+      "epoch": 1.1856,
+      "grad_norm": 0.2891118824481964,
+      "learning_rate": 3e-06,
+      "loss": 0.0728,
+      "policy/loss": 0.6593028177352096,
+      "step": 2220
+    },
+    {
+      "completion_length": 187.53125,
+      "epoch": 1.1861333333333333,
+      "grad_norm": 0.24903199076652527,
+      "learning_rate": 3e-06,
+      "loss": 0.0017,
+      "policy/loss": 0.5384423732757568,
+      "reward": 1.2708333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.5833333432674408,
+      "step": 2221
+    },
+    {
+      "epoch": 1.1866666666666668,
+      "grad_norm": 0.20168843865394592,
+      "learning_rate": 3e-06,
+      "loss": -0.1354,
+      "policy/loss": 0.00991543848067522,
+      "step": 2222
+    },
+    {
+      "epoch": 1.1872,
+      "grad_norm": 0.24493438005447388,
+      "learning_rate": 3e-06,
+      "loss": 0.001,
+      "policy/loss": 0.5403648614883423,
+      "step": 2223
+    },
+    {
+      "epoch": 1.1877333333333333,
+      "grad_norm": 0.20599794387817383,
+      "learning_rate": 3e-06,
+      "loss": -0.1361,
+      "policy/loss": 0.012761940248310566,
+      "step": 2224
+    },
+    {
+      "completion_length": 214.61458587646484,
+      "epoch": 1.1882666666666666,
+      "grad_norm": 0.2828269302845001,
+      "learning_rate": 3e-06,
+      "loss": 0.0332,
+      "policy/loss": 0.01727759838104248,
+      "reward": 1.7916667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333134651184,
+      "rewards/correctness_reward_func_math": 1.0833333432674408,
+      "step": 2225
+    },
+    {
+      "epoch": 1.1888,
+      "grad_norm": 0.3393276631832123,
+      "learning_rate": 3e-06,
+      "loss": 0.106,
+      "policy/loss": 0.7219237387180328,
+      "step": 2226
+    },
+    {
+      "epoch": 1.1893333333333334,
+      "grad_norm": 0.28760892152786255,
+      "learning_rate": 3e-06,
+      "loss": 0.0318,
+      "policy/loss": 0.010989919304847717,
+      "step": 2227
+    },
+    {
+      "epoch": 1.1898666666666666,
+      "grad_norm": 0.3363463878631592,
+      "learning_rate": 3e-06,
+      "loss": 0.1062,
+      "policy/loss": 0.7220421731472015,
+      "step": 2228
+    },
+    {
+      "completion_length": 224.6354217529297,
+      "epoch": 1.1904,
+      "grad_norm": 0.3136160373687744,
+      "learning_rate": 3e-06,
+      "loss": -0.1015,
+      "policy/loss": -0.05828594163536582,
+      "reward": 1.5156250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.8333333432674408,
+      "step": 2229
+    },
+    {
+      "epoch": 1.1909333333333334,
+      "grad_norm": 0.312659353017807,
+      "learning_rate": 3e-06,
+      "loss": -0.0822,
+      "policy/loss": -0.06773243483022995,
+      "step": 2230
+    },
+    {
+      "epoch": 1.1914666666666667,
+      "grad_norm": 0.3220772445201874,
+      "learning_rate": 3e-06,
+      "loss": -0.1012,
+      "policy/loss": -0.06046924861750114,
+      "step": 2231
+    },
+    {
+      "epoch": 1.192,
+      "grad_norm": 0.33633485436439514,
+      "learning_rate": 3e-06,
+      "loss": -0.083,
+      "policy/loss": -0.0669046678559786,
+      "step": 2232
+    },
+    {
+      "completion_length": 208.7291717529297,
+      "epoch": 1.1925333333333334,
+      "grad_norm": 0.29427069425582886,
+      "learning_rate": 3e-06,
+      "loss": 0.127,
+      "policy/loss": 0.7506137901640777,
+      "reward": 1.5625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.875,
+      "step": 2233
+    },
+    {
+      "epoch": 1.1930666666666667,
+      "grad_norm": 0.48521658778190613,
+      "learning_rate": 3e-06,
+      "loss": -0.0644,
+      "policy/loss": 0.26580535393284066,
+      "step": 2234
+    },
+    {
+      "epoch": 1.1936,
+      "grad_norm": 0.292625367641449,
+      "learning_rate": 3e-06,
+      "loss": 0.1249,
+      "policy/loss": 0.7469671781023699,
+      "step": 2235
+    },
+    {
+      "epoch": 1.1941333333333333,
+      "grad_norm": 0.2950514853000641,
+      "learning_rate": 3e-06,
+      "loss": -0.0656,
+      "policy/loss": 0.25984485973882343,
+      "step": 2236
+    },
+    {
+      "completion_length": 213.9791717529297,
+      "epoch": 1.1946666666666665,
+      "grad_norm": 0.3434818387031555,
+      "learning_rate": 3e-06,
+      "loss": -0.0503,
+      "policy/loss": -0.20670229196548462,
+      "reward": 1.7291666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333134651184,
+      "rewards/correctness_reward_func_math": 1.0208333134651184,
+      "step": 2237
+    },
+    {
+      "epoch": 1.1952,
+      "grad_norm": 0.31879135966300964,
+      "learning_rate": 3e-06,
+      "loss": -0.036,
+      "policy/loss": -0.30764415860176086,
+      "step": 2238
+    },
+    {
+      "epoch": 1.1957333333333333,
+      "grad_norm": 0.359730064868927,
+      "learning_rate": 3e-06,
+      "loss": -0.0512,
+      "policy/loss": -0.20832645893096924,
+      "step": 2239
+    },
+    {
+      "epoch": 1.1962666666666666,
+      "grad_norm": 0.28935477137565613,
+      "learning_rate": 3e-06,
+      "loss": -0.0371,
+      "policy/loss": -0.3075624853372574,
+      "step": 2240
+    },
+    {
+      "completion_length": 208.56250762939453,
+      "epoch": 1.1968,
+      "grad_norm": 0.27321016788482666,
+      "learning_rate": 3e-06,
+      "loss": 0.0538,
+      "policy/loss": -0.052582188631907556,
+      "reward": 1.2760416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 2241
+    },
+    {
+      "epoch": 1.1973333333333334,
+      "grad_norm": 0.3497752845287323,
+      "learning_rate": 3e-06,
+      "loss": 0.0013,
+      "policy/loss": -0.11835110945633431,
+      "step": 2242
+    },
+    {
+      "epoch": 1.1978666666666666,
+      "grad_norm": 0.2766272723674774,
+      "learning_rate": 3e-06,
+      "loss": 0.0531,
+      "policy/loss": -0.059553790063120005,
+      "step": 2243
+    },
+    {
+      "epoch": 1.1984,
+      "grad_norm": 0.3151801824569702,
+      "learning_rate": 3e-06,
+      "loss": -0.0007,
+      "policy/loss": -0.12723521899448365,
+      "step": 2244
+    },
+    {
+      "completion_length": 205.28125,
+      "epoch": 1.1989333333333334,
+      "grad_norm": 0.26176607608795166,
+      "learning_rate": 3e-06,
+      "loss": -0.0456,
+      "policy/loss": -0.05430779157400423,
+      "reward": 1.4270833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.7083333134651184,
+      "step": 2245
+    },
+    {
+      "epoch": 1.1994666666666667,
+      "grad_norm": 0.2870265543460846,
+      "learning_rate": 3e-06,
+      "loss": -0.0205,
+      "policy/loss": -0.004034063298469448,
+      "step": 2246
+    },
+    {
+      "epoch": 1.2,
+      "grad_norm": 0.2622152864933014,
+      "learning_rate": 3e-06,
+      "loss": -0.0466,
+      "policy/loss": -0.052769567077426416,
+      "step": 2247
+    },
+    {
+      "epoch": 1.2005333333333335,
+      "grad_norm": 0.2824257016181946,
+      "learning_rate": 3e-06,
+      "loss": -0.0199,
+      "policy/loss": -0.0022562930770391176,
+      "step": 2248
+    },
+    {
+      "completion_length": 211.90625762939453,
+      "epoch": 1.2010666666666667,
+      "grad_norm": 0.5556309819221497,
+      "learning_rate": 3e-06,
+      "loss": 0.0337,
+      "policy/loss": 1.5806271846940945e-07,
+      "reward": 1.5416666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 2249
+    },
+    {
+      "epoch": 1.2016,
+      "grad_norm": 0.3966769278049469,
+      "learning_rate": 3e-06,
+      "loss": 0.1124,
+      "policy/loss": 1.5975759026787273e-07,
+      "step": 2250
+    },
+    {
+      "epoch": 1.2021333333333333,
+      "grad_norm": 0.4763129651546478,
+      "learning_rate": 3e-06,
+      "loss": 0.0335,
+      "policy/loss": 1.5828864974309909e-07,
+      "step": 2251
+    },
+    {
+      "epoch": 1.2026666666666666,
+      "grad_norm": 0.36168181896209717,
+      "learning_rate": 3e-06,
+      "loss": 0.1129,
+      "policy/loss": 1.596334655573628e-07,
+      "step": 2252
+    },
+    {
+      "completion_length": 207.92708587646484,
+      "epoch": 1.2032,
+      "grad_norm": 0.2952141761779785,
+      "learning_rate": 3e-06,
+      "loss": 0.0975,
+      "policy/loss": 0.30618031229132825,
+      "reward": 1.1354167461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.59375,
+      "rewards/correctness_reward_func_math": 0.5416666716337204,
+      "step": 2253
+    },
+    {
+      "epoch": 1.2037333333333333,
+      "grad_norm": 0.3302564024925232,
+      "learning_rate": 3e-06,
+      "loss": -0.0571,
+      "policy/loss": -0.15834269400174605,
+      "step": 2254
+    },
+    {
+      "epoch": 1.2042666666666666,
+      "grad_norm": 0.2794647514820099,
+      "learning_rate": 3e-06,
+      "loss": 0.0964,
+      "policy/loss": 0.3064863524397552,
+      "step": 2255
+    },
+    {
+      "epoch": 1.2048,
+      "grad_norm": 0.32072967290878296,
+      "learning_rate": 3e-06,
+      "loss": -0.0582,
+      "policy/loss": -0.15980974307928264,
+      "step": 2256
+    },
+    {
+      "completion_length": 213.56250762939453,
+      "epoch": 1.2053333333333334,
+      "grad_norm": 0.33035194873809814,
+      "learning_rate": 3e-06,
+      "loss": 0.1952,
+      "policy/loss": 1.0504356851015473e-07,
+      "reward": 1.3645833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.7083333134651184,
+      "step": 2257
+    },
+    {
+      "epoch": 1.2058666666666666,
+      "grad_norm": 0.25941798090934753,
+      "learning_rate": 3e-06,
+      "loss": 0.0159,
+      "policy/loss": 1.1913201802826734e-07,
+      "step": 2258
+    },
+    {
+      "epoch": 1.2064,
+      "grad_norm": 0.3278129994869232,
+      "learning_rate": 3e-06,
+      "loss": 0.1937,
+      "policy/loss": 1.0488588486623485e-07,
+      "step": 2259
+    },
+    {
+      "epoch": 1.2069333333333334,
+      "grad_norm": 0.23757599294185638,
+      "learning_rate": 3e-06,
+      "loss": 0.0139,
+      "policy/loss": 1.193804841648216e-07,
+      "step": 2260
+    },
+    {
+      "completion_length": 173.9791717529297,
+      "epoch": 1.2074666666666667,
+      "grad_norm": 0.22863860428333282,
+      "learning_rate": 3e-06,
+      "loss": -0.0429,
+      "policy/loss": -0.3127324324613934,
+      "reward": 1.6354166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.9166666567325592,
+      "step": 2261
+    },
+    {
+      "epoch": 1.208,
+      "grad_norm": 0.2539094090461731,
+      "learning_rate": 3e-06,
+      "loss": 0.1166,
+      "policy/loss": 0.19180587530361137,
+      "step": 2262
+    },
+    {
+      "epoch": 1.2085333333333332,
+      "grad_norm": 0.22220255434513092,
+      "learning_rate": 3e-06,
+      "loss": -0.0443,
+      "policy/loss": -0.3122772021252054,
+      "step": 2263
+    },
+    {
+      "epoch": 1.2090666666666667,
+      "grad_norm": 0.23751281201839447,
+      "learning_rate": 3e-06,
+      "loss": 0.115,
+      "policy/loss": 0.18772990102181808,
+      "step": 2264
+    },
+    {
+      "completion_length": 221.89584350585938,
+      "epoch": 1.2096,
+      "grad_norm": 0.2622303366661072,
+      "learning_rate": 3e-06,
+      "loss": 0.1158,
+      "policy/loss": -0.30133581906557083,
+      "reward": 1.203125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.5208333432674408,
+      "step": 2265
+    },
+    {
+      "epoch": 1.2101333333333333,
+      "grad_norm": 0.22407843172550201,
+      "learning_rate": 3e-06,
+      "loss": -0.0981,
+      "policy/loss": -0.26394131779670715,
+      "step": 2266
+    },
+    {
+      "epoch": 1.2106666666666666,
+      "grad_norm": 0.2679302990436554,
+      "learning_rate": 3e-06,
+      "loss": 0.1145,
+      "policy/loss": -0.3040818050503731,
+      "step": 2267
+    },
+    {
+      "epoch": 1.2112,
+      "grad_norm": 0.2317013442516327,
+      "learning_rate": 3e-06,
+      "loss": -0.0982,
+      "policy/loss": -0.264116995036602,
+      "step": 2268
+    },
+    {
+      "completion_length": 195.15625,
+      "epoch": 1.2117333333333333,
+      "grad_norm": 0.24567312002182007,
+      "learning_rate": 3e-06,
+      "loss": -0.1155,
+      "policy/loss": -0.6081279756701008,
+      "reward": 1.9791667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 1.2291666269302368,
+      "step": 2269
+    },
+    {
+      "epoch": 1.2122666666666666,
+      "grad_norm": 0.24888132512569427,
+      "learning_rate": 3e-06,
+      "loss": 0.1538,
+      "policy/loss": -0.6257655820005219,
+      "step": 2270
+    },
+    {
+      "epoch": 1.2128,
+      "grad_norm": 0.2523675858974457,
+      "learning_rate": 3e-06,
+      "loss": -0.1168,
+      "policy/loss": -0.6092289331029512,
+      "step": 2271
+    },
+    {
+      "epoch": 1.2133333333333334,
+      "grad_norm": 0.2346271574497223,
+      "learning_rate": 3e-06,
+      "loss": 0.1531,
+      "policy/loss": -0.6233384211735107,
+      "step": 2272
+    },
+    {
+      "completion_length": 200.5104217529297,
+      "epoch": 1.2138666666666666,
+      "grad_norm": 0.36398619413375854,
+      "learning_rate": 3e-06,
+      "loss": 0.0364,
+      "policy/loss": 0.7710016242970781,
+      "reward": 1.6354166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.9166666567325592,
+      "step": 2273
+    },
+    {
+      "epoch": 1.2144,
+      "grad_norm": 0.31334277987480164,
+      "learning_rate": 3e-06,
+      "loss": -0.0962,
+      "policy/loss": 0.12832160838422624,
+      "step": 2274
+    },
+    {
+      "epoch": 1.2149333333333334,
+      "grad_norm": 0.32708051800727844,
+      "learning_rate": 3e-06,
+      "loss": 0.0351,
+      "policy/loss": 0.7730566732426816,
+      "step": 2275
+    },
+    {
+      "epoch": 1.2154666666666667,
+      "grad_norm": 0.28773027658462524,
+      "learning_rate": 3e-06,
+      "loss": -0.097,
+      "policy/loss": 0.1286328786973101,
+      "step": 2276
+    },
+    {
+      "completion_length": 209.90625,
+      "epoch": 1.216,
+      "grad_norm": 0.30413398146629333,
+      "learning_rate": 3e-06,
+      "loss": 0.0862,
+      "policy/loss": 0.011562853951410546,
+      "reward": 1.4375000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 2277
+    },
+    {
+      "epoch": 1.2165333333333332,
+      "grad_norm": 0.28782621026039124,
+      "learning_rate": 3e-06,
+      "loss": 0.1564,
+      "policy/loss": 0.18644250915912863,
+      "step": 2278
+    },
+    {
+      "epoch": 1.2170666666666667,
+      "grad_norm": 0.33971109986305237,
+      "learning_rate": 3e-06,
+      "loss": 0.0855,
+      "policy/loss": 0.012123891030377365,
+      "step": 2279
+    },
+    {
+      "epoch": 1.2176,
+      "grad_norm": 0.2875097393989563,
+      "learning_rate": 3e-06,
+      "loss": 0.1549,
+      "policy/loss": 0.18570324773350677,
+      "step": 2280
+    },
+    {
+      "completion_length": 197.20833587646484,
+      "epoch": 1.2181333333333333,
+      "grad_norm": 0.3155469000339508,
+      "learning_rate": 3e-06,
+      "loss": -0.0596,
+      "policy/loss": 5.682385406657886e-08,
+      "reward": 1.5833333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.8333333432674408,
+      "step": 2281
+    },
+    {
+      "epoch": 1.2186666666666666,
+      "grad_norm": 0.5038817524909973,
+      "learning_rate": 3e-06,
+      "loss": 0.1266,
+      "policy/loss": 6.517368689173964e-08,
+      "step": 2282
+    },
+    {
+      "epoch": 1.2192,
+      "grad_norm": 0.30923399329185486,
+      "learning_rate": 3e-06,
+      "loss": -0.0607,
+      "policy/loss": 5.680626458115512e-08,
+      "step": 2283
+    },
+    {
+      "epoch": 1.2197333333333333,
+      "grad_norm": 0.46444493532180786,
+      "learning_rate": 3e-06,
+      "loss": 0.127,
+      "policy/loss": 6.515010397833976e-08,
+      "step": 2284
+    },
+    {
+      "completion_length": 203.2604217529297,
+      "epoch": 1.2202666666666666,
+      "grad_norm": 0.21674488484859467,
+      "learning_rate": 3e-06,
+      "loss": 0.075,
+      "policy/loss": 0.019151348621733177,
+      "reward": 1.7083333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 1.0208333432674408,
+      "step": 2285
+    },
+    {
+      "epoch": 1.2208,
+      "grad_norm": 0.2305380403995514,
+      "learning_rate": 3e-06,
+      "loss": 0.1015,
+      "policy/loss": 0.04969726472074143,
+      "step": 2286
+    },
+    {
+      "epoch": 1.2213333333333334,
+      "grad_norm": 0.20058992505073547,
+      "learning_rate": 3e-06,
+      "loss": 0.0737,
+      "policy/loss": 0.01665177005533991,
+      "step": 2287
+    },
+    {
+      "epoch": 1.2218666666666667,
+      "grad_norm": 0.22454972565174103,
+      "learning_rate": 3e-06,
+      "loss": 0.1004,
+      "policy/loss": 0.049725804247273686,
+      "step": 2288
+    },
+    {
+      "completion_length": 220.3854217529297,
+      "epoch": 1.2224,
+      "grad_norm": 0.35818424820899963,
+      "learning_rate": 3e-06,
+      "loss": 0.0896,
+      "policy/loss": 1.1885216721907454e-07,
+      "reward": 1.1145833432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.59375,
+      "rewards/correctness_reward_func_math": 0.5208333283662796,
+      "step": 2289
+    },
+    {
+      "epoch": 1.2229333333333334,
+      "grad_norm": 0.37288743257522583,
+      "learning_rate": 3e-06,
+      "loss": 0.0409,
+      "policy/loss": 1.0342939305019172e-07,
+      "step": 2290
+    },
+    {
+      "epoch": 1.2234666666666667,
+      "grad_norm": 0.3853507936000824,
+      "learning_rate": 3e-06,
+      "loss": 0.0878,
+      "policy/loss": 1.1869729732438827e-07,
+      "step": 2291
+    },
+    {
+      "epoch": 1.224,
+      "grad_norm": 0.362736314535141,
+      "learning_rate": 3e-06,
+      "loss": 0.0381,
+      "policy/loss": 1.0349761225825205e-07,
+      "step": 2292
+    },
+    {
+      "completion_length": 202.8229217529297,
+      "epoch": 1.2245333333333333,
+      "grad_norm": 0.24891284108161926,
+      "learning_rate": 3e-06,
+      "loss": -0.082,
+      "policy/loss": 0.011851871194458852,
+      "reward": 1.5625000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 2293
+    },
+    {
+      "epoch": 1.2250666666666667,
+      "grad_norm": 0.23000890016555786,
+      "learning_rate": 3e-06,
+      "loss": 0.0508,
+      "policy/loss": 0.012799223781534863,
+      "step": 2294
+    },
+    {
+      "epoch": 1.2256,
+      "grad_norm": 0.24358424544334412,
+      "learning_rate": 3e-06,
+      "loss": -0.083,
+      "policy/loss": 0.010747220221627884,
+      "step": 2295
+    },
+    {
+      "epoch": 1.2261333333333333,
+      "grad_norm": 0.24139955639839172,
+      "learning_rate": 3e-06,
+      "loss": 0.0487,
+      "policy/loss": 0.014982802681256402,
+      "step": 2296
+    },
+    {
+      "completion_length": 195.89584350585938,
+      "epoch": 1.2266666666666666,
+      "grad_norm": 0.2679109573364258,
+      "learning_rate": 3e-06,
+      "loss": 0.0879,
+      "policy/loss": 7.722841033341865e-08,
+      "reward": 1.65625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 1.0,
+      "step": 2297
+    },
+    {
+      "epoch": 1.2272,
+      "grad_norm": 0.29927241802215576,
+      "learning_rate": 3e-06,
+      "loss": 0.0441,
+      "policy/loss": 6.032123778254572e-08,
+      "step": 2298
+    },
+    {
+      "epoch": 1.2277333333333333,
+      "grad_norm": 0.26938027143478394,
+      "learning_rate": 3e-06,
+      "loss": 0.0869,
+      "policy/loss": 7.718829309055764e-08,
+      "step": 2299
+    },
+    {
+      "epoch": 1.2282666666666666,
+      "grad_norm": 0.27282413840293884,
+      "learning_rate": 3e-06,
+      "loss": 0.0424,
+      "policy/loss": 6.028406929203811e-08,
+      "step": 2300
+    },
+    {
+      "completion_length": 205.28125762939453,
+      "epoch": 1.2288000000000001,
+      "grad_norm": 0.5657669305801392,
+      "learning_rate": 3e-06,
+      "loss": 0.0949,
+      "policy/loss": 9.431891001554504e-08,
+      "reward": 1.2916666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.5416666567325592,
+      "step": 2301
+    },
+    {
+      "epoch": 1.2293333333333334,
+      "grad_norm": 0.22467860579490662,
+      "learning_rate": 3e-06,
+      "loss": 0.0097,
+      "policy/loss": 1.0090291269193585e-07,
+      "step": 2302
+    },
+    {
+      "epoch": 1.2298666666666667,
+      "grad_norm": 0.28043556213378906,
+      "learning_rate": 3e-06,
+      "loss": 0.0958,
+      "policy/loss": 9.43586684343245e-08,
+      "step": 2303
+    },
+    {
+      "epoch": 1.2304,
+      "grad_norm": 0.24245385825634003,
+      "learning_rate": 3e-06,
+      "loss": 0.0095,
+      "policy/loss": 1.0086014867738413e-07,
+      "step": 2304
+    },
+    {
+      "completion_length": 208.55208587646484,
+      "epoch": 1.2309333333333332,
+      "grad_norm": 0.2067108303308487,
+      "learning_rate": 3e-06,
+      "loss": 0.0815,
+      "policy/loss": 2.9378774968336074e-08,
+      "reward": 1.0833333432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.3958333283662796,
+      "step": 2305
+    },
+    {
+      "epoch": 1.2314666666666667,
+      "grad_norm": 0.2683736979961395,
+      "learning_rate": 3e-06,
+      "loss": 0.0108,
+      "policy/loss": 4.3495099077972554e-08,
+      "step": 2306
+    },
+    {
+      "epoch": 1.232,
+      "grad_norm": 0.20954500138759613,
+      "learning_rate": 3e-06,
+      "loss": 0.0809,
+      "policy/loss": 2.9356526098922586e-08,
+      "step": 2307
+    },
+    {
+      "epoch": 1.2325333333333333,
+      "grad_norm": 0.25297507643699646,
+      "learning_rate": 3e-06,
+      "loss": 0.0096,
+      "policy/loss": 4.349922377855364e-08,
+      "step": 2308
+    },
+    {
+      "completion_length": 200.39583587646484,
+      "epoch": 1.2330666666666668,
+      "grad_norm": 0.26269006729125977,
+      "learning_rate": 3e-06,
+      "loss": 0.038,
+      "policy/loss": -0.24672781063495286,
+      "reward": 1.4791667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 2309
+    },
+    {
+      "epoch": 1.2336,
+      "grad_norm": 0.28406327962875366,
+      "learning_rate": 3e-06,
+      "loss": -0.0566,
+      "policy/loss": -0.21131258013330267,
+      "step": 2310
+    },
+    {
+      "epoch": 1.2341333333333333,
+      "grad_norm": 0.26296132802963257,
+      "learning_rate": 3e-06,
+      "loss": 0.0374,
+      "policy/loss": -0.24688673152819973,
+      "step": 2311
+    },
+    {
+      "epoch": 1.2346666666666666,
+      "grad_norm": 0.26601889729499817,
+      "learning_rate": 3e-06,
+      "loss": -0.0569,
+      "policy/loss": -0.20985243023559264,
+      "step": 2312
+    },
+    {
+      "completion_length": 192.98958587646484,
+      "epoch": 1.2352,
+      "grad_norm": 0.30640876293182373,
+      "learning_rate": 3e-06,
+      "loss": -0.0272,
+      "policy/loss": 1.3181915647919595e-07,
+      "reward": 1.7708333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 1.0208333134651184,
+      "step": 2313
+    },
+    {
+      "epoch": 1.2357333333333334,
+      "grad_norm": 0.3039802610874176,
+      "learning_rate": 3e-06,
+      "loss": -0.1114,
+      "policy/loss": 1.4026887740214988e-07,
+      "step": 2314
+    },
+    {
+      "epoch": 1.2362666666666666,
+      "grad_norm": 0.34454938769340515,
+      "learning_rate": 3e-06,
+      "loss": -0.0267,
+      "policy/loss": 1.3190626901860014e-07,
+      "step": 2315
+    },
+    {
+      "epoch": 1.2368000000000001,
+      "grad_norm": 0.39287304878234863,
+      "learning_rate": 3e-06,
+      "loss": -0.1135,
+      "policy/loss": 1.4036252338200939e-07,
+      "step": 2316
+    },
+    {
+      "completion_length": 212.1041717529297,
+      "epoch": 1.2373333333333334,
+      "grad_norm": 0.33453282713890076,
+      "learning_rate": 3e-06,
+      "loss": 0.0302,
+      "policy/loss": -0.3475853744894266,
+      "reward": 1.4427083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 2317
+    },
+    {
+      "epoch": 1.2378666666666667,
+      "grad_norm": 0.310959130525589,
+      "learning_rate": 3e-06,
+      "loss": -0.046,
+      "policy/loss": -0.22490724921226501,
+      "step": 2318
+    },
+    {
+      "epoch": 1.2384,
+      "grad_norm": 0.3094745874404907,
+      "learning_rate": 3e-06,
+      "loss": 0.0278,
+      "policy/loss": -0.34905526228249073,
+      "step": 2319
+    },
+    {
+      "epoch": 1.2389333333333332,
+      "grad_norm": 0.2768138349056244,
+      "learning_rate": 3e-06,
+      "loss": -0.0478,
+      "policy/loss": -0.22641977667808533,
+      "step": 2320
+    },
+    {
+      "completion_length": 210.80208587646484,
+      "epoch": 1.2394666666666667,
+      "grad_norm": 0.31262171268463135,
+      "learning_rate": 3e-06,
+      "loss": -0.0135,
+      "policy/loss": 0.3303765028034604,
+      "reward": 1.46875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 2321
+    },
+    {
+      "epoch": 1.24,
+      "grad_norm": 0.305929034948349,
+      "learning_rate": 3e-06,
+      "loss": -0.0854,
+      "policy/loss": 0.10222248711340498,
+      "step": 2322
+    },
+    {
+      "epoch": 1.2405333333333333,
+      "grad_norm": 0.27600154280662537,
+      "learning_rate": 3e-06,
+      "loss": -0.0133,
+      "policy/loss": 0.3324038058101131,
+      "step": 2323
+    },
+    {
+      "epoch": 1.2410666666666668,
+      "grad_norm": 0.347572386264801,
+      "learning_rate": 3e-06,
+      "loss": -0.0867,
+      "policy/loss": 0.09914496521605898,
+      "step": 2324
+    },
+    {
+      "completion_length": 185.18750762939453,
+      "epoch": 1.2416,
+      "grad_norm": 0.4299040734767914,
+      "learning_rate": 3e-06,
+      "loss": 0.0818,
+      "policy/loss": 0.05086229778229878,
+      "reward": 1.46875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6354166567325592,
+      "rewards/correctness_reward_func_math": 0.8333333432674408,
+      "step": 2325
+    },
+    {
+      "epoch": 1.2421333333333333,
+      "grad_norm": 0.5109900832176208,
+      "learning_rate": 3e-06,
+      "loss": 0.2005,
+      "policy/loss": -0.08910650951886723,
+      "step": 2326
+    },
+    {
+      "epoch": 1.2426666666666666,
+      "grad_norm": 0.4256366789340973,
+      "learning_rate": 3e-06,
+      "loss": 0.0795,
+      "policy/loss": 0.050738528745828404,
+      "step": 2327
+    },
+    {
+      "epoch": 1.2432,
+      "grad_norm": 0.41070297360420227,
+      "learning_rate": 3e-06,
+      "loss": 0.1992,
+      "policy/loss": -0.0891971235354454,
+      "step": 2328
+    },
+    {
+      "completion_length": 204.71875,
+      "epoch": 1.2437333333333334,
+      "grad_norm": 0.2697182595729828,
+      "learning_rate": 3e-06,
+      "loss": -0.1058,
+      "policy/loss": 1.525174440075716e-07,
+      "reward": 1.5677083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.8541666567325592,
+      "step": 2329
+    },
+    {
+      "epoch": 1.2442666666666666,
+      "grad_norm": 0.32857853174209595,
+      "learning_rate": 3e-06,
+      "loss": -0.0677,
+      "policy/loss": 1.650707517342198e-07,
+      "step": 2330
+    },
+    {
+      "epoch": 1.2448,
+      "grad_norm": 0.2599409818649292,
+      "learning_rate": 3e-06,
+      "loss": -0.1082,
+      "policy/loss": 1.5261393571108783e-07,
+      "step": 2331
+    },
+    {
+      "epoch": 1.2453333333333334,
+      "grad_norm": 0.31497520208358765,
+      "learning_rate": 3e-06,
+      "loss": -0.0691,
+      "policy/loss": 1.6516650447329084e-07,
+      "step": 2332
+    },
+    {
+      "completion_length": 194.96875762939453,
+      "epoch": 1.2458666666666667,
+      "grad_norm": 0.3082866668701172,
+      "learning_rate": 3e-06,
+      "loss": 0.0248,
+      "policy/loss": -0.15567653346806765,
+      "reward": 1.6354166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.9166666567325592,
+      "step": 2333
+    },
+    {
+      "epoch": 1.2464,
+      "grad_norm": 0.33698219060897827,
+      "learning_rate": 3e-06,
+      "loss": 0.1421,
+      "policy/loss": 0.49545852839946747,
+      "step": 2334
+    },
+    {
+      "epoch": 1.2469333333333332,
+      "grad_norm": 0.30848294496536255,
+      "learning_rate": 3e-06,
+      "loss": 0.0232,
+      "policy/loss": -0.15692751668393612,
+      "step": 2335
+    },
+    {
+      "epoch": 1.2474666666666667,
+      "grad_norm": 0.33765459060668945,
+      "learning_rate": 3e-06,
+      "loss": 0.1414,
+      "policy/loss": 0.49687062203884125,
+      "step": 2336
+    },
+    {
+      "completion_length": 201.1666717529297,
+      "epoch": 1.248,
+      "grad_norm": 0.2950269877910614,
+      "learning_rate": 3e-06,
+      "loss": 0.0211,
+      "policy/loss": 1.5100582828608822e-07,
+      "reward": 1.6041667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.8541666567325592,
+      "step": 2337
+    },
+    {
+      "epoch": 1.2485333333333333,
+      "grad_norm": 0.2536742389202118,
+      "learning_rate": 3e-06,
+      "loss": -0.1863,
+      "policy/loss": 1.396222799598945e-07,
+      "step": 2338
+    },
+    {
+      "epoch": 1.2490666666666668,
+      "grad_norm": 0.2789474427700043,
+      "learning_rate": 3e-06,
+      "loss": 0.0199,
+      "policy/loss": 1.5085797855363126e-07,
+      "step": 2339
+    },
+    {
+      "epoch": 1.2496,
+      "grad_norm": 0.2853201627731323,
+      "learning_rate": 3e-06,
+      "loss": -0.1869,
+      "policy/loss": 1.3962550582391486e-07,
+      "step": 2340
+    },
+    {
+      "completion_length": 194.06250762939453,
+      "epoch": 1.2501333333333333,
+      "grad_norm": 0.2611059844493866,
+      "learning_rate": 3e-06,
+      "loss": 0.0965,
+      "policy/loss": 9.842681336635906e-08,
+      "reward": 1.2968750298023224,
+      "rewards/boxed_and_answer_tags_format_reward": 0.5885416716337204,
+      "rewards/correctness_reward_func_math": 0.7083333283662796,
+      "step": 2341
+    },
+    {
+      "epoch": 1.2506666666666666,
+      "grad_norm": 0.39422598481178284,
+      "learning_rate": 3e-06,
+      "loss": -0.0205,
+      "policy/loss": 9.384387311683895e-08,
+      "step": 2342
+    },
+    {
+      "epoch": 1.2511999999999999,
+      "grad_norm": 0.25900447368621826,
+      "learning_rate": 3e-06,
+      "loss": 0.0955,
+      "policy/loss": 9.835475722752562e-08,
+      "step": 2343
+    },
+    {
+      "epoch": 1.2517333333333334,
+      "grad_norm": 0.36099085211753845,
+      "learning_rate": 3e-06,
+      "loss": -0.0242,
+      "policy/loss": 9.37627042674194e-08,
+      "step": 2344
+    },
+    {
+      "completion_length": 199.17709350585938,
+      "epoch": 1.2522666666666666,
+      "grad_norm": 0.31740835309028625,
+      "learning_rate": 3e-06,
+      "loss": -0.0816,
+      "policy/loss": -0.01350894570350647,
+      "reward": 1.4791666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 2345
+    },
+    {
+      "epoch": 1.2528000000000001,
+      "grad_norm": 0.34207016229629517,
+      "learning_rate": 3e-06,
+      "loss": 0.0364,
+      "policy/loss": 0.09860067162662745,
+      "step": 2346
+    },
+    {
+      "epoch": 1.2533333333333334,
+      "grad_norm": 0.31878629326820374,
+      "learning_rate": 3e-06,
+      "loss": -0.0825,
+      "policy/loss": -0.019929498434066772,
+      "step": 2347
+    },
+    {
+      "epoch": 1.2538666666666667,
+      "grad_norm": 0.34836071729660034,
+      "learning_rate": 3e-06,
+      "loss": 0.0342,
+      "policy/loss": 0.09078513877466321,
+      "step": 2348
+    },
+    {
+      "completion_length": 203.6354217529297,
+      "epoch": 1.2544,
+      "grad_norm": 0.2432815134525299,
+      "learning_rate": 3e-06,
+      "loss": 0.05,
+      "policy/loss": 0.44265829075398244,
+      "reward": 1.8437500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 1.125,
+      "step": 2349
+    },
+    {
+      "epoch": 1.2549333333333332,
+      "grad_norm": 0.34775999188423157,
+      "learning_rate": 3e-06,
+      "loss": 0.115,
+      "policy/loss": -0.09585408432377918,
+      "step": 2350
+    },
+    {
+      "epoch": 1.2554666666666667,
+      "grad_norm": 0.2558012306690216,
+      "learning_rate": 3e-06,
+      "loss": 0.0487,
+      "policy/loss": 0.43898199552963035,
+      "step": 2351
+    },
+    {
+      "epoch": 1.256,
+      "grad_norm": 0.3124086558818817,
+      "learning_rate": 3e-06,
+      "loss": 0.114,
+      "policy/loss": -0.09756569868086018,
+      "step": 2352
+    },
+    {
+      "completion_length": 206.34375,
+      "epoch": 1.2565333333333333,
+      "grad_norm": 0.3162170350551605,
+      "learning_rate": 3e-06,
+      "loss": -0.0824,
+      "policy/loss": 0.0,
+      "reward": 1.40625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 2353
+    },
+    {
+      "epoch": 1.2570666666666668,
+      "grad_norm": 0.327260822057724,
+      "learning_rate": 3e-06,
+      "loss": -0.1136,
+      "policy/loss": 0.0,
+      "step": 2354
+    },
+    {
+      "epoch": 1.2576,
+      "grad_norm": 0.31556186079978943,
+      "learning_rate": 3e-06,
+      "loss": -0.0841,
+      "policy/loss": 0.0,
+      "step": 2355
+    },
+    {
+      "epoch": 1.2581333333333333,
+      "grad_norm": 0.3231678605079651,
+      "learning_rate": 3e-06,
+      "loss": -0.1152,
+      "policy/loss": 0.0,
+      "step": 2356
+    },
+    {
+      "completion_length": 204.2916717529297,
+      "epoch": 1.2586666666666666,
+      "grad_norm": 0.32848867774009705,
+      "learning_rate": 3e-06,
+      "loss": 0.0963,
+      "policy/loss": 0.16893902646488357,
+      "reward": 1.1979166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.4791666567325592,
+      "step": 2357
+    },
+    {
+      "epoch": 1.2591999999999999,
+      "grad_norm": 0.33033323287963867,
+      "learning_rate": 3e-06,
+      "loss": 0.1211,
+      "policy/loss": 0.4954755439021099,
+      "step": 2358
+    },
+    {
+      "epoch": 1.2597333333333334,
+      "grad_norm": 0.32962629199028015,
+      "learning_rate": 3e-06,
+      "loss": 0.0957,
+      "policy/loss": 0.17017221683536832,
+      "step": 2359
+    },
+    {
+      "epoch": 1.2602666666666666,
+      "grad_norm": 0.33097442984580994,
+      "learning_rate": 3e-06,
+      "loss": 0.1183,
+      "policy/loss": 0.49108497634761505,
+      "step": 2360
+    },
+    {
+      "completion_length": 198.89584350585938,
+      "epoch": 1.2608,
+      "grad_norm": 0.3176988959312439,
+      "learning_rate": 3e-06,
+      "loss": -0.0684,
+      "policy/loss": -0.8767070546746254,
+      "reward": 1.0989583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.4166666716337204,
+      "step": 2361
+    },
+    {
+      "epoch": 1.2613333333333334,
+      "grad_norm": 0.28102144598960876,
+      "learning_rate": 3e-06,
+      "loss": -0.1617,
+      "policy/loss": -0.1648842692375183,
+      "step": 2362
+    },
+    {
+      "epoch": 1.2618666666666667,
+      "grad_norm": 0.3068493902683258,
+      "learning_rate": 3e-06,
+      "loss": -0.0711,
+      "policy/loss": -0.8850802518427372,
+      "step": 2363
+    },
+    {
+      "epoch": 1.2624,
+      "grad_norm": 0.2763919532299042,
+      "learning_rate": 3e-06,
+      "loss": -0.1621,
+      "policy/loss": -0.16540493071079254,
+      "step": 2364
+    },
+    {
+      "completion_length": 199.4375,
+      "epoch": 1.2629333333333332,
+      "grad_norm": 0.2637392282485962,
+      "learning_rate": 3e-06,
+      "loss": -0.0454,
+      "policy/loss": -0.11509272292894224,
+      "reward": 1.34375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 2365
+    },
+    {
+      "epoch": 1.2634666666666667,
+      "grad_norm": 0.3295654356479645,
+      "learning_rate": 3e-06,
+      "loss": -0.0315,
+      "policy/loss": -0.4510951648731272,
+      "step": 2366
+    },
+    {
+      "epoch": 1.264,
+      "grad_norm": 0.26858365535736084,
+      "learning_rate": 3e-06,
+      "loss": -0.047,
+      "policy/loss": -0.11337805378250465,
+      "step": 2367
+    },
+    {
+      "epoch": 1.2645333333333333,
+      "grad_norm": 0.29552125930786133,
+      "learning_rate": 3e-06,
+      "loss": -0.0326,
+      "policy/loss": -0.4512186954689845,
+      "step": 2368
+    },
+    {
+      "completion_length": 196.98958587646484,
+      "epoch": 1.2650666666666668,
+      "grad_norm": 0.3295383155345917,
+      "learning_rate": 3e-06,
+      "loss": -0.1333,
+      "policy/loss": 0.09593792607751794,
+      "reward": 1.708333432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.958333358168602,
+      "step": 2369
+    },
+    {
+      "epoch": 1.2656,
+      "grad_norm": 0.24892808496952057,
+      "learning_rate": 3e-06,
+      "loss": -0.0339,
+      "policy/loss": 0.06196785444534214,
+      "step": 2370
+    },
+    {
+      "epoch": 1.2661333333333333,
+      "grad_norm": 0.32095345854759216,
+      "learning_rate": 3e-06,
+      "loss": -0.1357,
+      "policy/loss": 0.0924143752265465,
+      "step": 2371
+    },
+    {
+      "epoch": 1.2666666666666666,
+      "grad_norm": 0.24044223129749298,
+      "learning_rate": 3e-06,
+      "loss": -0.0348,
+      "policy/loss": 0.06168962733500649,
+      "step": 2372
+    },
+    {
+      "completion_length": 187.05209350585938,
+      "epoch": 1.2671999999999999,
+      "grad_norm": 0.2311103492975235,
+      "learning_rate": 3e-06,
+      "loss": 0.2273,
+      "policy/loss": -0.007934526192132196,
+      "reward": 1.6354166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.59375,
+      "rewards/correctness_reward_func_math": 1.0416666567325592,
+      "step": 2373
+    },
+    {
+      "epoch": 1.2677333333333334,
+      "grad_norm": 0.24935388565063477,
+      "learning_rate": 3e-06,
+      "loss": 0.1858,
+      "policy/loss": 0.33765976831459454,
+      "step": 2374
+    },
+    {
+      "epoch": 1.2682666666666667,
+      "grad_norm": 0.21646635234355927,
+      "learning_rate": 3e-06,
+      "loss": 0.2265,
+      "policy/loss": -0.008759030775221888,
+      "step": 2375
+    },
+    {
+      "epoch": 1.2688,
+      "grad_norm": 0.2440306693315506,
+      "learning_rate": 3e-06,
+      "loss": 0.1843,
+      "policy/loss": 0.3339707772655416,
+      "step": 2376
+    },
+    {
+      "completion_length": 196.5416717529297,
+      "epoch": 1.2693333333333334,
+      "grad_norm": 0.3594660460948944,
+      "learning_rate": 3e-06,
+      "loss": 0.0834,
+      "policy/loss": 0.2643061801791191,
+      "reward": 1.5729166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8541666567325592,
+      "step": 2377
+    },
+    {
+      "epoch": 1.2698666666666667,
+      "grad_norm": 0.38957875967025757,
+      "learning_rate": 3e-06,
+      "loss": 0.0901,
+      "policy/loss": 0.47149835526943207,
+      "step": 2378
+    },
+    {
+      "epoch": 1.2704,
+      "grad_norm": 0.3386372923851013,
+      "learning_rate": 3e-06,
+      "loss": 0.081,
+      "policy/loss": 0.2594783678650856,
+      "step": 2379
+    },
+    {
+      "epoch": 1.2709333333333332,
+      "grad_norm": 0.37060633301734924,
+      "learning_rate": 3e-06,
+      "loss": 0.0892,
+      "policy/loss": 0.4705689251422882,
+      "step": 2380
+    },
+    {
+      "completion_length": 206.14583587646484,
+      "epoch": 1.2714666666666667,
+      "grad_norm": 0.280294269323349,
+      "learning_rate": 3e-06,
+      "loss": -0.0667,
+      "policy/loss": -0.09074809417046481,
+      "reward": 1.3854166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6666666716337204,
+      "step": 2381
+    },
+    {
+      "epoch": 1.272,
+      "grad_norm": 0.27054375410079956,
+      "learning_rate": 3e-06,
+      "loss": 0.1086,
+      "policy/loss": 0.3538542766560937,
+      "step": 2382
+    },
+    {
+      "epoch": 1.2725333333333333,
+      "grad_norm": 0.28501129150390625,
+      "learning_rate": 3e-06,
+      "loss": -0.0673,
+      "policy/loss": -0.09070231785788962,
+      "step": 2383
+    },
+    {
+      "epoch": 1.2730666666666668,
+      "grad_norm": 0.2675800323486328,
+      "learning_rate": 3e-06,
+      "loss": 0.1068,
+      "policy/loss": 0.3522251326430563,
+      "step": 2384
+    },
+    {
+      "completion_length": 180.7291717529297,
+      "epoch": 1.2736,
+      "grad_norm": 0.366440087556839,
+      "learning_rate": 3e-06,
+      "loss": 0.1406,
+      "policy/loss": 0.20768187586450892,
+      "reward": 1.5312500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 2385
+    },
+    {
+      "epoch": 1.2741333333333333,
+      "grad_norm": 0.3943886160850525,
+      "learning_rate": 3e-06,
+      "loss": 0.1997,
+      "policy/loss": 0.8948598416584659,
+      "step": 2386
+    },
+    {
+      "epoch": 1.2746666666666666,
+      "grad_norm": 0.3560185730457306,
+      "learning_rate": 3e-06,
+      "loss": 0.1388,
+      "policy/loss": 0.20533433218442454,
+      "step": 2387
+    },
+    {
+      "epoch": 1.2752,
+      "grad_norm": 0.6283380389213562,
+      "learning_rate": 3e-06,
+      "loss": 0.1968,
+      "policy/loss": 0.8918188723362412,
+      "step": 2388
+    },
+    {
+      "completion_length": 184.09375762939453,
+      "epoch": 1.2757333333333334,
+      "grad_norm": 0.2994515001773834,
+      "learning_rate": 3e-06,
+      "loss": -0.003,
+      "policy/loss": 0.21809179335832596,
+      "reward": 1.4375000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 2389
+    },
+    {
+      "epoch": 1.2762666666666667,
+      "grad_norm": 0.29477643966674805,
+      "learning_rate": 3e-06,
+      "loss": -0.0104,
+      "policy/loss": 0.2803344987332821,
+      "step": 2390
+    },
+    {
+      "epoch": 1.2768,
+      "grad_norm": 0.3107386827468872,
+      "learning_rate": 3e-06,
+      "loss": -0.0048,
+      "policy/loss": 0.21349888760596514,
+      "step": 2391
+    },
+    {
+      "epoch": 1.2773333333333334,
+      "grad_norm": 0.3213185966014862,
+      "learning_rate": 3e-06,
+      "loss": -0.0118,
+      "policy/loss": 0.27773576602339745,
+      "step": 2392
+    },
+    {
+      "completion_length": 168.89584350585938,
+      "epoch": 1.2778666666666667,
+      "grad_norm": 0.2928459644317627,
+      "learning_rate": 3e-06,
+      "loss": -0.0672,
+      "policy/loss": 0.09179239217098001,
+      "reward": 1.875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 1.125,
+      "step": 2393
+    },
+    {
+      "epoch": 1.2784,
+      "grad_norm": 0.40558791160583496,
+      "learning_rate": 3e-06,
+      "loss": 0.0341,
+      "policy/loss": 0.2365035611366011,
+      "step": 2394
+    },
+    {
+      "epoch": 1.2789333333333333,
+      "grad_norm": 0.3014681041240692,
+      "learning_rate": 3e-06,
+      "loss": -0.0679,
+      "policy/loss": 0.09134064853169122,
+      "step": 2395
+    },
+    {
+      "epoch": 1.2794666666666665,
+      "grad_norm": 0.6366944313049316,
+      "learning_rate": 3e-06,
+      "loss": 0.0321,
+      "policy/loss": 0.23625387720514368,
+      "step": 2396
+    },
+    {
+      "completion_length": 199.8229217529297,
+      "epoch": 1.28,
+      "grad_norm": 0.2845044732093811,
+      "learning_rate": 3e-06,
+      "loss": -0.0662,
+      "policy/loss": 1.1321986193024713e-07,
+      "reward": 1.7760416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 1.0625,
+      "step": 2397
+    },
+    {
+      "epoch": 1.2805333333333333,
+      "grad_norm": 0.28162163496017456,
+      "learning_rate": 3e-06,
+      "loss": 0.0559,
+      "policy/loss": 1.2039712871114716e-07,
+      "step": 2398
+    },
+    {
+      "epoch": 1.2810666666666668,
+      "grad_norm": 0.2947237193584442,
+      "learning_rate": 3e-06,
+      "loss": -0.0683,
+      "policy/loss": 1.1332626215221353e-07,
+      "step": 2399
+    },
+    {
+      "epoch": 1.2816,
+      "grad_norm": 0.2757442593574524,
+      "learning_rate": 3e-06,
+      "loss": 0.055,
+      "policy/loss": 1.2023981454944987e-07,
+      "step": 2400
+    },
+    {
+      "completion_length": 198.0729217529297,
+      "epoch": 1.2821333333333333,
+      "grad_norm": 0.3158217668533325,
+      "learning_rate": 3e-06,
+      "loss": 0.0915,
+      "policy/loss": 9.597898120716764e-08,
+      "reward": 1.4583333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7708333134651184,
+      "step": 2401
+    },
+    {
+      "epoch": 1.2826666666666666,
+      "grad_norm": 0.2659715414047241,
+      "learning_rate": 3e-06,
+      "loss": 0.0782,
+      "policy/loss": 9.042095427957975e-08,
+      "step": 2402
+    },
+    {
+      "epoch": 1.2832,
+      "grad_norm": 0.32197409868240356,
+      "learning_rate": 3e-06,
+      "loss": 0.0899,
+      "policy/loss": 9.594125316425561e-08,
+      "step": 2403
+    },
+    {
+      "epoch": 1.2837333333333334,
+      "grad_norm": 0.2420838475227356,
+      "learning_rate": 3e-06,
+      "loss": 0.0765,
+      "policy/loss": 9.04685979463693e-08,
+      "step": 2404
+    },
+    {
+      "completion_length": 188.1979217529297,
+      "epoch": 1.2842666666666667,
+      "grad_norm": 0.20697100460529327,
+      "learning_rate": 3e-06,
+      "loss": -0.1507,
+      "policy/loss": 1.481173725892404e-07,
+      "reward": 1.2708333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.5833333544433117,
+      "step": 2405
+    },
+    {
+      "epoch": 1.2848,
+      "grad_norm": 0.2420910894870758,
+      "learning_rate": 3e-06,
+      "loss": -0.0508,
+      "policy/loss": 1.4755257282672574e-07,
+      "step": 2406
+    },
+    {
+      "epoch": 1.2853333333333334,
+      "grad_norm": 0.20840488374233246,
+      "learning_rate": 3e-06,
+      "loss": -0.1509,
+      "policy/loss": 1.4807404014050007e-07,
+      "step": 2407
+    },
+    {
+      "epoch": 1.2858666666666667,
+      "grad_norm": 0.26204755902290344,
+      "learning_rate": 3e-06,
+      "loss": -0.0525,
+      "policy/loss": 1.4761024047516003e-07,
+      "step": 2408
+    },
+    {
+      "completion_length": 184.64583587646484,
+      "epoch": 1.2864,
+      "grad_norm": 0.2967803478240967,
+      "learning_rate": 3e-06,
+      "loss": -0.0788,
+      "policy/loss": 4.569694489475751e-08,
+      "reward": 1.4062500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6875000149011612,
+      "step": 2409
+    },
+    {
+      "epoch": 1.2869333333333333,
+      "grad_norm": 0.2991659343242645,
+      "learning_rate": 3e-06,
+      "loss": -0.176,
+      "policy/loss": 3.265434589394545e-08,
+      "step": 2410
+    },
+    {
+      "epoch": 1.2874666666666665,
+      "grad_norm": 0.29713666439056396,
+      "learning_rate": 3e-06,
+      "loss": -0.0796,
+      "policy/loss": 4.5685935035066905e-08,
+      "step": 2411
+    },
+    {
+      "epoch": 1.288,
+      "grad_norm": 0.30846691131591797,
+      "learning_rate": 3e-06,
+      "loss": -0.1772,
+      "policy/loss": 3.265659387352571e-08,
+      "step": 2412
+    },
+    {
+      "completion_length": 192.05208587646484,
+      "epoch": 1.2885333333333333,
+      "grad_norm": 0.300060898065567,
+      "learning_rate": 3e-06,
+      "loss": -0.0676,
+      "policy/loss": -0.23984361720964031,
+      "reward": 1.2916666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.5416666567325592,
+      "step": 2413
+    },
+    {
+      "epoch": 1.2890666666666668,
+      "grad_norm": 0.4418605864048004,
+      "learning_rate": 3e-06,
+      "loss": -0.0009,
+      "policy/loss": -0.06260539336838633,
+      "step": 2414
+    },
+    {
+      "epoch": 1.2896,
+      "grad_norm": 0.29900574684143066,
+      "learning_rate": 3e-06,
+      "loss": -0.0698,
+      "policy/loss": -0.2376845432651109,
+      "step": 2415
+    },
+    {
+      "epoch": 1.2901333333333334,
+      "grad_norm": 0.4454765319824219,
+      "learning_rate": 3e-06,
+      "loss": -0.003,
+      "policy/loss": -0.06298116334912152,
+      "step": 2416
+    },
+    {
+      "completion_length": 200.1666717529297,
+      "epoch": 1.2906666666666666,
+      "grad_norm": 0.2658022344112396,
+      "learning_rate": 3e-06,
+      "loss": -0.0143,
+      "policy/loss": 0.12202106136828661,
+      "reward": 1.4739583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 2417
+    },
+    {
+      "epoch": 1.2912,
+      "grad_norm": 0.27238234877586365,
+      "learning_rate": 3e-06,
+      "loss": -0.0152,
+      "policy/loss": -0.22379267401993275,
+      "step": 2418
+    },
+    {
+      "epoch": 1.2917333333333334,
+      "grad_norm": 0.2585465908050537,
+      "learning_rate": 3e-06,
+      "loss": -0.016,
+      "policy/loss": 0.12181963585317135,
+      "step": 2419
+    },
+    {
+      "epoch": 1.2922666666666667,
+      "grad_norm": 0.26277002692222595,
+      "learning_rate": 3e-06,
+      "loss": -0.0166,
+      "policy/loss": -0.2240750417113304,
+      "step": 2420
+    },
+    {
+      "completion_length": 218.75000762939453,
+      "epoch": 1.2928,
+      "grad_norm": 0.48113852739334106,
+      "learning_rate": 3e-06,
+      "loss": -0.0937,
+      "policy/loss": 6.521173290252591e-08,
+      "reward": 1.2500000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.5625,
+      "step": 2421
+    },
+    {
+      "epoch": 1.2933333333333334,
+      "grad_norm": 0.27617594599723816,
+      "learning_rate": 3e-06,
+      "loss": -0.0028,
+      "policy/loss": 6.857387369052503e-08,
+      "step": 2422
+    },
+    {
+      "epoch": 1.2938666666666667,
+      "grad_norm": 0.40892285108566284,
+      "learning_rate": 3e-06,
+      "loss": -0.0957,
+      "policy/loss": 6.52538911793954e-08,
+      "step": 2423
+    },
+    {
+      "epoch": 1.2944,
+      "grad_norm": 0.32359182834625244,
+      "learning_rate": 3e-06,
+      "loss": -0.0042,
+      "policy/loss": 6.852133971335661e-08,
+      "step": 2424
+    },
+    {
+      "completion_length": 198.7604217529297,
+      "epoch": 1.2949333333333333,
+      "grad_norm": 0.30989354848861694,
+      "learning_rate": 3e-06,
+      "loss": -0.1563,
+      "policy/loss": 8.546539831399969e-08,
+      "reward": 1.5625000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 2425
+    },
+    {
+      "epoch": 1.2954666666666665,
+      "grad_norm": 0.2971349060535431,
+      "learning_rate": 3e-06,
+      "loss": -0.1023,
+      "policy/loss": 8.635760373465473e-08,
+      "step": 2426
+    },
+    {
+      "epoch": 1.296,
+      "grad_norm": 0.3101601302623749,
+      "learning_rate": 3e-06,
+      "loss": -0.1579,
+      "policy/loss": 8.546938268239046e-08,
+      "step": 2427
+    },
+    {
+      "epoch": 1.2965333333333333,
+      "grad_norm": 0.31514209508895874,
+      "learning_rate": 3e-06,
+      "loss": -0.105,
+      "policy/loss": 8.646797766687087e-08,
+      "step": 2428
+    },
+    {
+      "completion_length": 198.6354217529297,
+      "epoch": 1.2970666666666666,
+      "grad_norm": 0.31083959341049194,
+      "learning_rate": 3e-06,
+      "loss": 0.371,
+      "policy/loss": 0.03035178780555725,
+      "reward": 1.7916667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6458333432674408,
+      "rewards/correctness_reward_func_math": 1.1458333134651184,
+      "step": 2429
+    },
+    {
+      "epoch": 1.2976,
+      "grad_norm": 0.31467971205711365,
+      "learning_rate": 3e-06,
+      "loss": 0.1688,
+      "policy/loss": 0.42276251316070557,
+      "step": 2430
+    },
+    {
+      "epoch": 1.2981333333333334,
+      "grad_norm": 0.3134637176990509,
+      "learning_rate": 3e-06,
+      "loss": 0.3704,
+      "policy/loss": 0.03154098987579346,
+      "step": 2431
+    },
+    {
+      "epoch": 1.2986666666666666,
+      "grad_norm": 0.3217748999595642,
+      "learning_rate": 3e-06,
+      "loss": 0.1678,
+      "policy/loss": 0.42149747908115387,
+      "step": 2432
+    },
+    {
+      "completion_length": 207.71875762939453,
+      "epoch": 1.2992,
+      "grad_norm": 0.30527597665786743,
+      "learning_rate": 3e-06,
+      "loss": -0.115,
+      "policy/loss": -0.5886734022746296,
+      "reward": 1.4322916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 2433
+    },
+    {
+      "epoch": 1.2997333333333334,
+      "grad_norm": 0.3150326907634735,
+      "learning_rate": 3e-06,
+      "loss": -0.0753,
+      "policy/loss": -0.14177112992536678,
+      "step": 2434
+    },
+    {
+      "epoch": 1.3002666666666667,
+      "grad_norm": 0.3421999216079712,
+      "learning_rate": 3e-06,
+      "loss": -0.1164,
+      "policy/loss": -0.5911133780118263,
+      "step": 2435
+    },
+    {
+      "epoch": 1.3008,
+      "grad_norm": 0.33123698830604553,
+      "learning_rate": 3e-06,
+      "loss": -0.0756,
+      "policy/loss": -0.14262058544269962,
+      "step": 2436
+    },
+    {
+      "completion_length": 202.28125762939453,
+      "epoch": 1.3013333333333335,
+      "grad_norm": 0.32603803277015686,
+      "learning_rate": 3e-06,
+      "loss": 0.0442,
+      "policy/loss": -0.46600822221918037,
+      "reward": 1.2239583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.5416666716337204,
+      "step": 2437
+    },
+    {
+      "epoch": 1.3018666666666667,
+      "grad_norm": 0.34371665120124817,
+      "learning_rate": 3e-06,
+      "loss": 0.1192,
+      "policy/loss": -0.5542747655650437,
+      "step": 2438
+    },
+    {
+      "epoch": 1.3024,
+      "grad_norm": 0.32394781708717346,
+      "learning_rate": 3e-06,
+      "loss": 0.0434,
+      "policy/loss": -0.467090851201835,
+      "step": 2439
+    },
+    {
+      "epoch": 1.3029333333333333,
+      "grad_norm": 0.3779315650463104,
+      "learning_rate": 3e-06,
+      "loss": 0.1167,
+      "policy/loss": -0.5601634660788122,
+      "step": 2440
+    },
+    {
+      "completion_length": 200.5104217529297,
+      "epoch": 1.3034666666666666,
+      "grad_norm": 0.1829119175672531,
+      "learning_rate": 3e-06,
+      "loss": -0.0075,
+      "policy/loss": 7.005147750760443e-08,
+      "reward": 0.9635416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6927083432674408,
+      "rewards/correctness_reward_func_math": 0.2708333283662796,
+      "step": 2441
+    },
+    {
+      "epoch": 1.304,
+      "grad_norm": 0.25530245900154114,
+      "learning_rate": 3e-06,
+      "loss": 0.1151,
+      "policy/loss": 6.42944755213648e-08,
+      "step": 2442
+    },
+    {
+      "epoch": 1.3045333333333333,
+      "grad_norm": 0.18792738020420074,
+      "learning_rate": 3e-06,
+      "loss": -0.0077,
+      "policy/loss": 7.009169422644845e-08,
+      "step": 2443
+    },
+    {
+      "epoch": 1.3050666666666666,
+      "grad_norm": 0.21456167101860046,
+      "learning_rate": 3e-06,
+      "loss": 0.1141,
+      "policy/loss": 6.408697572624078e-08,
+      "step": 2444
+    },
+    {
+      "completion_length": 211.84375,
+      "epoch": 1.3056,
+      "grad_norm": 0.22989121079444885,
+      "learning_rate": 3e-06,
+      "loss": -0.035,
+      "policy/loss": 0.02331965139770631,
+      "reward": 1.2708333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.5833333283662796,
+      "step": 2445
+    },
+    {
+      "epoch": 1.3061333333333334,
+      "grad_norm": 0.2981398105621338,
+      "learning_rate": 3e-06,
+      "loss": -0.0593,
+      "policy/loss": -0.16341052804697398,
+      "step": 2446
+    },
+    {
+      "epoch": 1.3066666666666666,
+      "grad_norm": 0.22538787126541138,
+      "learning_rate": 3e-06,
+      "loss": -0.0362,
+      "policy/loss": 0.02208913774907728,
+      "step": 2447
+    },
+    {
+      "epoch": 1.3072,
+      "grad_norm": 0.2872328758239746,
+      "learning_rate": 3e-06,
+      "loss": -0.0597,
+      "policy/loss": -0.1630174503652384,
+      "step": 2448
+    },
+    {
+      "completion_length": 200.1666717529297,
+      "epoch": 1.3077333333333334,
+      "grad_norm": 0.2963619530200958,
+      "learning_rate": 3e-06,
+      "loss": -0.165,
+      "policy/loss": -0.6295258240462189,
+      "reward": 1.6822916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 1.0,
+      "step": 2449
+    },
+    {
+      "epoch": 1.3082666666666667,
+      "grad_norm": 0.3495224118232727,
+      "learning_rate": 3e-06,
+      "loss": -0.1913,
+      "policy/loss": -0.9259071737333358,
+      "step": 2450
+    },
+    {
+      "epoch": 1.3088,
+      "grad_norm": 0.29473963379859924,
+      "learning_rate": 3e-06,
+      "loss": -0.1664,
+      "policy/loss": -0.6334047750540748,
+      "step": 2451
+    },
+    {
+      "epoch": 1.3093333333333335,
+      "grad_norm": 0.3749459683895111,
+      "learning_rate": 3e-06,
+      "loss": -0.1927,
+      "policy/loss": -0.9316560416362876,
+      "step": 2452
+    },
+    {
+      "completion_length": 197.15625762939453,
+      "epoch": 1.3098666666666667,
+      "grad_norm": 0.33580657839775085,
+      "learning_rate": 3e-06,
+      "loss": -0.0116,
+      "policy/loss": 0.11009947955608368,
+      "reward": 1.5989583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.640625,
+      "rewards/correctness_reward_func_math": 0.9583333432674408,
+      "step": 2453
+    },
+    {
+      "epoch": 1.3104,
+      "grad_norm": 0.3592522442340851,
+      "learning_rate": 3e-06,
+      "loss": 0.1109,
+      "policy/loss": 0.5855512171983719,
+      "step": 2454
+    },
+    {
+      "epoch": 1.3109333333333333,
+      "grad_norm": 0.33993539214134216,
+      "learning_rate": 3e-06,
+      "loss": -0.0124,
+      "policy/loss": 0.1101926863193512,
+      "step": 2455
+    },
+    {
+      "epoch": 1.3114666666666666,
+      "grad_norm": 0.33256855607032776,
+      "learning_rate": 3e-06,
+      "loss": 0.108,
+      "policy/loss": 0.5842734426259995,
+      "step": 2456
+    },
+    {
+      "completion_length": 183.9166717529297,
+      "epoch": 1.312,
+      "grad_norm": 0.3796480596065521,
+      "learning_rate": 3e-06,
+      "loss": -0.0358,
+      "policy/loss": 0.5154255951375006,
+      "reward": 1.359375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 2457
+    },
+    {
+      "epoch": 1.3125333333333333,
+      "grad_norm": 0.2896620035171509,
+      "learning_rate": 3e-06,
+      "loss": 0.0153,
+      "policy/loss": 0.377778365387897,
+      "step": 2458
+    },
+    {
+      "epoch": 1.3130666666666666,
+      "grad_norm": 0.3667716383934021,
+      "learning_rate": 3e-06,
+      "loss": -0.036,
+      "policy/loss": 0.5135297501716707,
+      "step": 2459
+    },
+    {
+      "epoch": 1.3136,
+      "grad_norm": 0.2939019203186035,
+      "learning_rate": 3e-06,
+      "loss": 0.0137,
+      "policy/loss": 0.376651539596903,
+      "step": 2460
+    },
+    {
+      "completion_length": 212.68750762939453,
+      "epoch": 1.3141333333333334,
+      "grad_norm": 0.3496158719062805,
+      "learning_rate": 3e-06,
+      "loss": 0.124,
+      "policy/loss": 1.4134831261003455e-07,
+      "reward": 1.2500000298023224,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.6249999850988388,
+      "step": 2461
+    },
+    {
+      "epoch": 1.3146666666666667,
+      "grad_norm": 0.38349616527557373,
+      "learning_rate": 3e-06,
+      "loss": 0.2107,
+      "policy/loss": 1.5346730819487675e-07,
+      "step": 2462
+    },
+    {
+      "epoch": 1.3152,
+      "grad_norm": 0.3358713686466217,
+      "learning_rate": 3e-06,
+      "loss": 0.1219,
+      "policy/loss": 1.414055681436821e-07,
+      "step": 2463
+    },
+    {
+      "epoch": 1.3157333333333332,
+      "grad_norm": 0.4544832706451416,
+      "learning_rate": 3e-06,
+      "loss": 0.2098,
+      "policy/loss": 1.535574760680447e-07,
+      "step": 2464
+    },
+    {
+      "completion_length": 192.71875,
+      "epoch": 1.3162666666666667,
+      "grad_norm": 0.3226768672466278,
+      "learning_rate": 3e-06,
+      "loss": 0.0392,
+      "policy/loss": 1.1277952762547194,
+      "reward": 1.5364583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6197916567325592,
+      "rewards/correctness_reward_func_math": 0.9166666567325592,
+      "step": 2465
+    },
+    {
+      "epoch": 1.3168,
+      "grad_norm": 0.26943379640579224,
+      "learning_rate": 3e-06,
+      "loss": 0.0478,
+      "policy/loss": 0.566904768025374,
+      "step": 2466
+    },
+    {
+      "epoch": 1.3173333333333335,
+      "grad_norm": 0.324917197227478,
+      "learning_rate": 3e-06,
+      "loss": 0.0379,
+      "policy/loss": 1.1202843758495895,
+      "step": 2467
+    },
+    {
+      "epoch": 1.3178666666666667,
+      "grad_norm": 0.26934120059013367,
+      "learning_rate": 3e-06,
+      "loss": 0.0459,
+      "policy/loss": 0.5627342906088586,
+      "step": 2468
+    },
+    {
+      "completion_length": 217.7291717529297,
+      "epoch": 1.3184,
+      "grad_norm": 0.3935249149799347,
+      "learning_rate": 3e-06,
+      "loss": 0.0084,
+      "policy/loss": 0.2291645811836034,
+      "reward": 1.1666666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.4791666716337204,
+      "step": 2469
+    },
+    {
+      "epoch": 1.3189333333333333,
+      "grad_norm": 0.5418949723243713,
+      "learning_rate": 3e-06,
+      "loss": 0.0747,
+      "policy/loss": 0.3857283505762812,
+      "step": 2470
+    },
+    {
+      "epoch": 1.3194666666666666,
+      "grad_norm": 0.38324373960494995,
+      "learning_rate": 3e-06,
+      "loss": 0.0085,
+      "policy/loss": 0.232329960306787,
+      "step": 2471
+    },
+    {
+      "epoch": 1.32,
+      "grad_norm": 0.4801650047302246,
+      "learning_rate": 3e-06,
+      "loss": 0.0721,
+      "policy/loss": 0.3793888601626776,
+      "step": 2472
+    },
+    {
+      "completion_length": 195.73959350585938,
+      "epoch": 1.3205333333333333,
+      "grad_norm": 0.29419198632240295,
+      "learning_rate": 3e-06,
+      "loss": -0.0262,
+      "policy/loss": -1.1543950736522675,
+      "reward": 1.6718750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.9583333134651184,
+      "step": 2473
+    },
+    {
+      "epoch": 1.3210666666666666,
+      "grad_norm": 0.25514867901802063,
+      "learning_rate": 3e-06,
+      "loss": -0.0605,
+      "policy/loss": -0.5363017022609711,
+      "step": 2474
+    },
+    {
+      "epoch": 1.3216,
+      "grad_norm": 0.28030523657798767,
+      "learning_rate": 3e-06,
+      "loss": -0.0269,
+      "policy/loss": -1.156757265329361,
+      "step": 2475
+    },
+    {
+      "epoch": 1.3221333333333334,
+      "grad_norm": 0.24419336020946503,
+      "learning_rate": 3e-06,
+      "loss": -0.062,
+      "policy/loss": -0.535103902220726,
+      "step": 2476
+    },
+    {
+      "completion_length": 195.4479217529297,
+      "epoch": 1.3226666666666667,
+      "grad_norm": 0.37785500288009644,
+      "learning_rate": 3e-06,
+      "loss": 0.186,
+      "policy/loss": 0.3718823753792613,
+      "reward": 1.1458333432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.4583333283662796,
+      "step": 2477
+    },
+    {
+      "epoch": 1.3232,
+      "grad_norm": 0.33318382501602173,
+      "learning_rate": 3e-06,
+      "loss": -0.0917,
+      "policy/loss": 0.011563311346339589,
+      "step": 2478
+    },
+    {
+      "epoch": 1.3237333333333332,
+      "grad_norm": 0.35618504881858826,
+      "learning_rate": 3e-06,
+      "loss": 0.1847,
+      "policy/loss": 0.3689165376945738,
+      "step": 2479
+    },
+    {
+      "epoch": 1.3242666666666667,
+      "grad_norm": 0.3768264949321747,
+      "learning_rate": 3e-06,
+      "loss": -0.0937,
+      "policy/loss": 0.00835324181901953,
+      "step": 2480
+    },
+    {
+      "completion_length": 212.30208587646484,
+      "epoch": 1.3248,
+      "grad_norm": 0.28157150745391846,
+      "learning_rate": 3e-06,
+      "loss": -0.0779,
+      "policy/loss": -0.019139699290839474,
+      "reward": 1.6927083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 1.0416666567325592,
+      "step": 2481
+    },
+    {
+      "epoch": 1.3253333333333333,
+      "grad_norm": 0.25106772780418396,
+      "learning_rate": 3e-06,
+      "loss": 0.1417,
+      "policy/loss": 0.06814665704418132,
+      "step": 2482
+    },
+    {
+      "epoch": 1.3258666666666667,
+      "grad_norm": 0.2630786895751953,
+      "learning_rate": 3e-06,
+      "loss": -0.0788,
+      "policy/loss": -0.019557581998967066,
+      "step": 2483
+    },
+    {
+      "epoch": 1.3264,
+      "grad_norm": 0.22142016887664795,
+      "learning_rate": 3e-06,
+      "loss": 0.1408,
+      "policy/loss": 0.06948620422537033,
+      "step": 2484
+    },
+    {
+      "completion_length": 211.20833587646484,
+      "epoch": 1.3269333333333333,
+      "grad_norm": 0.21023082733154297,
+      "learning_rate": 3e-06,
+      "loss": -0.0094,
+      "policy/loss": -0.05468692258000374,
+      "reward": 1.0729166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.3541666567325592,
+      "step": 2485
+    },
+    {
+      "epoch": 1.3274666666666666,
+      "grad_norm": 0.21885527670383453,
+      "learning_rate": 3e-06,
+      "loss": -0.0183,
+      "policy/loss": 0.07790738344192505,
+      "step": 2486
+    },
+    {
+      "epoch": 1.328,
+      "grad_norm": 0.20121917128562927,
+      "learning_rate": 3e-06,
+      "loss": -0.0096,
+      "policy/loss": -0.05506939813494682,
+      "step": 2487
+    },
+    {
+      "epoch": 1.3285333333333333,
+      "grad_norm": 0.2512589693069458,
+      "learning_rate": 3e-06,
+      "loss": -0.0192,
+      "policy/loss": 0.07867888361215591,
+      "step": 2488
+    },
+    {
+      "completion_length": 206.78125762939453,
+      "epoch": 1.3290666666666666,
+      "grad_norm": 0.29935386776924133,
+      "learning_rate": 3e-06,
+      "loss": 0.0525,
+      "policy/loss": -0.26596634835004807,
+      "reward": 1.5104166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.8541666567325592,
+      "step": 2489
+    },
+    {
+      "epoch": 1.3296000000000001,
+      "grad_norm": 0.3184269964694977,
+      "learning_rate": 3e-06,
+      "loss": 0.1711,
+      "policy/loss": 0.2568120136857033,
+      "step": 2490
+    },
+    {
+      "epoch": 1.3301333333333334,
+      "grad_norm": 0.28586506843566895,
+      "learning_rate": 3e-06,
+      "loss": 0.0521,
+      "policy/loss": -0.26698122173547745,
+      "step": 2491
+    },
+    {
+      "epoch": 1.3306666666666667,
+      "grad_norm": 0.35003671050071716,
+      "learning_rate": 3e-06,
+      "loss": 0.1694,
+      "policy/loss": 0.25229762494564056,
+      "step": 2492
+    },
+    {
+      "completion_length": 208.6666717529297,
+      "epoch": 1.3312,
+      "grad_norm": 0.4074956476688385,
+      "learning_rate": 3e-06,
+      "loss": 0.0558,
+      "policy/loss": 9.182216587078074e-08,
+      "reward": 1.0833333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.3333333432674408,
+      "step": 2493
+    },
+    {
+      "epoch": 1.3317333333333332,
+      "grad_norm": 0.26277589797973633,
+      "learning_rate": 3e-06,
+      "loss": 0.0588,
+      "policy/loss": 1.0481022627573111e-07,
+      "step": 2494
+    },
+    {
+      "epoch": 1.3322666666666667,
+      "grad_norm": 0.5374290943145752,
+      "learning_rate": 3e-06,
+      "loss": 0.0558,
+      "policy/loss": 9.161893643749863e-08,
+      "step": 2495
+    },
+    {
+      "epoch": 1.3328,
+      "grad_norm": 0.2761056423187256,
+      "learning_rate": 3e-06,
+      "loss": 0.0581,
+      "policy/loss": 1.0483136492211997e-07,
+      "step": 2496
+    },
+    {
+      "completion_length": 184.75000762939453,
+      "epoch": 1.3333333333333333,
+      "grad_norm": 0.3651328384876251,
+      "learning_rate": 3e-06,
+      "loss": -0.1586,
+      "policy/loss": 0.23164304308086336,
+      "reward": 1.4947916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.8125,
+      "step": 2497
+    },
+    {
+      "epoch": 1.3338666666666668,
+      "grad_norm": 0.27503159642219543,
+      "learning_rate": 3e-06,
+      "loss": 0.0942,
+      "policy/loss": -0.18097375589662867,
+      "step": 2498
+    },
+    {
+      "epoch": 1.3344,
+      "grad_norm": 0.3529713451862335,
+      "learning_rate": 3e-06,
+      "loss": -0.1589,
+      "policy/loss": 0.2342139255561939,
+      "step": 2499
+    },
+    {
+      "epoch": 1.3349333333333333,
+      "grad_norm": 0.26590409874916077,
+      "learning_rate": 3e-06,
+      "loss": 0.0935,
+      "policy/loss": -0.1805903639141917,
+      "step": 2500
+    },
+    {
+      "completion_length": 205.3229217529297,
+      "epoch": 1.3354666666666666,
+      "grad_norm": 0.30370208621025085,
+      "learning_rate": 3e-06,
+      "loss": 0.0493,
+      "policy/loss": -0.051653635410943366,
+      "reward": 1.7604166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 1.0416666865348816,
+      "step": 2501
+    },
+    {
+      "epoch": 1.336,
+      "grad_norm": 0.3710249066352844,
+      "learning_rate": 3e-06,
+      "loss": 0.0761,
+      "policy/loss": 0.25720930039465273,
+      "step": 2502
+    },
+    {
+      "epoch": 1.3365333333333334,
+      "grad_norm": 0.3214002549648285,
+      "learning_rate": 3e-06,
+      "loss": 0.0485,
+      "policy/loss": -0.05198973490102077,
+      "step": 2503
+    },
+    {
+      "epoch": 1.3370666666666666,
+      "grad_norm": 0.3582456707954407,
+      "learning_rate": 3e-06,
+      "loss": 0.0742,
+      "policy/loss": 0.2521750623119452,
+      "step": 2504
+    },
+    {
+      "completion_length": 208.27083587646484,
+      "epoch": 1.3376000000000001,
+      "grad_norm": 0.627336859703064,
+      "learning_rate": 3e-06,
+      "loss": -0.0327,
+      "policy/loss": -0.47016897836832783,
+      "reward": 1.7031250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 1.0208333432674408,
+      "step": 2505
+    },
+    {
+      "epoch": 1.3381333333333334,
+      "grad_norm": 0.30435001850128174,
+      "learning_rate": 3e-06,
+      "loss": 0.153,
+      "policy/loss": 0.22190494373028713,
+      "step": 2506
+    },
+    {
+      "epoch": 1.3386666666666667,
+      "grad_norm": 0.310902863740921,
+      "learning_rate": 3e-06,
+      "loss": -0.0331,
+      "policy/loss": -0.4697751706156943,
+      "step": 2507
+    },
+    {
+      "epoch": 1.3392,
+      "grad_norm": 0.29129496216773987,
+      "learning_rate": 3e-06,
+      "loss": 0.1514,
+      "policy/loss": 0.21862948968735552,
+      "step": 2508
+    },
+    {
+      "completion_length": 216.39584350585938,
+      "epoch": 1.3397333333333332,
+      "grad_norm": 0.25039926171302795,
+      "learning_rate": 3e-06,
+      "loss": 0.0379,
+      "policy/loss": 0.48205293826870843,
+      "reward": 1.8750001192092896,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 1.1250000298023224,
+      "step": 2509
+    },
+    {
+      "epoch": 1.3402666666666667,
+      "grad_norm": 0.2638635039329529,
+      "learning_rate": 3e-06,
+      "loss": -0.1087,
+      "policy/loss": -0.31165681178184457,
+      "step": 2510
+    },
+    {
+      "epoch": 1.3408,
+      "grad_norm": 0.2515105605125427,
+      "learning_rate": 3e-06,
+      "loss": 0.0365,
+      "policy/loss": 0.4774983088090039,
+      "step": 2511
+    },
+    {
+      "epoch": 1.3413333333333333,
+      "grad_norm": 0.2647181749343872,
+      "learning_rate": 3e-06,
+      "loss": -0.1102,
+      "policy/loss": -0.3146430045648785,
+      "step": 2512
+    },
+    {
+      "completion_length": 215.90625762939453,
+      "epoch": 1.3418666666666668,
+      "grad_norm": 0.29243388772010803,
+      "learning_rate": 3e-06,
+      "loss": -0.1082,
+      "policy/loss": 1.2087709322372575e-07,
+      "reward": 1.1718750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.4583333283662796,
+      "step": 2513
+    },
+    {
+      "epoch": 1.3424,
+      "grad_norm": 0.36713144183158875,
+      "learning_rate": 3e-06,
+      "loss": -0.0069,
+      "policy/loss": 1.0353726764833482e-07,
+      "step": 2514
+    },
+    {
+      "epoch": 1.3429333333333333,
+      "grad_norm": 0.2900346517562866,
+      "learning_rate": 3e-06,
+      "loss": -0.1096,
+      "policy/loss": 1.2076905520075343e-07,
+      "step": 2515
+    },
+    {
+      "epoch": 1.3434666666666666,
+      "grad_norm": 0.36733025312423706,
+      "learning_rate": 3e-06,
+      "loss": -0.0072,
+      "policy/loss": 1.0361403823822002e-07,
+      "step": 2516
+    },
+    {
+      "completion_length": 198.27083587646484,
+      "epoch": 1.3439999999999999,
+      "grad_norm": 0.5390698909759521,
+      "learning_rate": 3e-06,
+      "loss": 0.0504,
+      "policy/loss": 9.399842326729413e-08,
+      "reward": 1.421875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.7083333283662796,
+      "step": 2517
+    },
+    {
+      "epoch": 1.3445333333333334,
+      "grad_norm": 0.26896488666534424,
+      "learning_rate": 3e-06,
+      "loss": 0.1354,
+      "policy/loss": 1.0381558546157521e-07,
+      "step": 2518
+    },
+    {
+      "epoch": 1.3450666666666666,
+      "grad_norm": 0.24797187745571136,
+      "learning_rate": 3e-06,
+      "loss": 0.0499,
+      "policy/loss": 9.393875899377235e-08,
+      "step": 2519
+    },
+    {
+      "epoch": 1.3456000000000001,
+      "grad_norm": 0.2872030735015869,
+      "learning_rate": 3e-06,
+      "loss": 0.1344,
+      "policy/loss": 1.0377466530542279e-07,
+      "step": 2520
+    },
+    {
+      "completion_length": 188.9375,
+      "epoch": 1.3461333333333334,
+      "grad_norm": 0.32145029306411743,
+      "learning_rate": 3e-06,
+      "loss": -0.1737,
+      "policy/loss": -0.24802645295858383,
+      "reward": 1.7812501192092896,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 1.0625,
+      "step": 2521
+    },
+    {
+      "epoch": 1.3466666666666667,
+      "grad_norm": 0.34083274006843567,
+      "learning_rate": 3e-06,
+      "loss": -0.1645,
+      "policy/loss": -0.13733054418116808,
+      "step": 2522
+    },
+    {
+      "epoch": 1.3472,
+      "grad_norm": 0.3319898247718811,
+      "learning_rate": 3e-06,
+      "loss": -0.1749,
+      "policy/loss": -0.25107239186763763,
+      "step": 2523
+    },
+    {
+      "epoch": 1.3477333333333332,
+      "grad_norm": 0.37128975987434387,
+      "learning_rate": 3e-06,
+      "loss": -0.1665,
+      "policy/loss": -0.13759192801080644,
+      "step": 2524
+    },
+    {
+      "completion_length": 203.7916717529297,
+      "epoch": 1.3482666666666667,
+      "grad_norm": 0.17933519184589386,
+      "learning_rate": 3e-06,
+      "loss": 0.1307,
+      "policy/loss": 9.477483686737287e-09,
+      "reward": 0.9895833432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.53125,
+      "rewards/correctness_reward_func_math": 0.4583333283662796,
+      "step": 2525
+    },
+    {
+      "epoch": 1.3488,
+      "grad_norm": 0.1972203403711319,
+      "learning_rate": 3e-06,
+      "loss": -0.0139,
+      "policy/loss": 9.676058176921742e-09,
+      "step": 2526
+    },
+    {
+      "epoch": 1.3493333333333333,
+      "grad_norm": 0.18752387166023254,
+      "learning_rate": 3e-06,
+      "loss": 0.1295,
+      "policy/loss": 9.473590800723741e-09,
+      "step": 2527
+    },
+    {
+      "epoch": 1.3498666666666668,
+      "grad_norm": 0.1864018738269806,
+      "learning_rate": 3e-06,
+      "loss": -0.015,
+      "policy/loss": 9.672588952014394e-09,
+      "step": 2528
+    },
+    {
+      "completion_length": 208.55209350585938,
+      "epoch": 1.3504,
+      "grad_norm": 0.3146488666534424,
+      "learning_rate": 3e-06,
+      "loss": 0.1117,
+      "policy/loss": 1.3895425254872862e-07,
+      "reward": 1.1145833432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.3958333283662796,
+      "step": 2529
+    },
+    {
+      "epoch": 1.3509333333333333,
+      "grad_norm": 0.271640419960022,
+      "learning_rate": 3e-06,
+      "loss": 0.1025,
+      "policy/loss": 1.4212483634423734e-07,
+      "step": 2530
+    },
+    {
+      "epoch": 1.3514666666666666,
+      "grad_norm": 0.3183027505874634,
+      "learning_rate": 3e-06,
+      "loss": 0.1084,
+      "policy/loss": 1.3870072734789574e-07,
+      "step": 2531
+    },
+    {
+      "epoch": 1.3519999999999999,
+      "grad_norm": 0.2826581597328186,
+      "learning_rate": 3e-06,
+      "loss": 0.1008,
+      "policy/loss": 1.4207162735146994e-07,
+      "step": 2532
+    },
+    {
+      "completion_length": 192.7604217529297,
+      "epoch": 1.3525333333333334,
+      "grad_norm": 0.23475028574466705,
+      "learning_rate": 3e-06,
+      "loss": -0.118,
+      "policy/loss": 1.276543315498202e-07,
+      "reward": 1.4375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 2533
+    },
+    {
+      "epoch": 1.3530666666666666,
+      "grad_norm": 0.27437037229537964,
+      "learning_rate": 3e-06,
+      "loss": 0.0362,
+      "policy/loss": 1.293199396457112e-07,
+      "step": 2534
+    },
+    {
+      "epoch": 1.3536000000000001,
+      "grad_norm": 0.24092476069927216,
+      "learning_rate": 3e-06,
+      "loss": -0.1192,
+      "policy/loss": 1.2749506339559957e-07,
+      "step": 2535
+    },
+    {
+      "epoch": 1.3541333333333334,
+      "grad_norm": 0.26687538623809814,
+      "learning_rate": 3e-06,
+      "loss": 0.0351,
+      "policy/loss": 1.292542783914996e-07,
+      "step": 2536
+    },
+    {
+      "completion_length": 211.78125762939453,
+      "epoch": 1.3546666666666667,
+      "grad_norm": 0.353071004152298,
+      "learning_rate": 3e-06,
+      "loss": 0.0826,
+      "policy/loss": -0.04216649012398932,
+      "reward": 1.375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 2537
+    },
+    {
+      "epoch": 1.3552,
+      "grad_norm": 0.31436148285865784,
+      "learning_rate": 3e-06,
+      "loss": 0.1092,
+      "policy/loss": 0.34998061832434146,
+      "step": 2538
+    },
+    {
+      "epoch": 1.3557333333333332,
+      "grad_norm": 0.3251684010028839,
+      "learning_rate": 3e-06,
+      "loss": 0.0807,
+      "policy/loss": -0.04479647423013944,
+      "step": 2539
+    },
+    {
+      "epoch": 1.3562666666666667,
+      "grad_norm": 0.32493603229522705,
+      "learning_rate": 3e-06,
+      "loss": 0.1082,
+      "policy/loss": 0.3464536327330272,
+      "step": 2540
+    },
+    {
+      "completion_length": 191.625,
+      "epoch": 1.3568,
+      "grad_norm": 0.32833462953567505,
+      "learning_rate": 3e-06,
+      "loss": -0.12,
+      "policy/loss": -0.41984617710113525,
+      "reward": 1.3802083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.6666666567325592,
+      "step": 2541
+    },
+    {
+      "epoch": 1.3573333333333333,
+      "grad_norm": 0.33843016624450684,
+      "learning_rate": 3e-06,
+      "loss": 0.0172,
+      "policy/loss": -0.0846124216914177,
+      "step": 2542
+    },
+    {
+      "epoch": 1.3578666666666668,
+      "grad_norm": 0.3254735469818115,
+      "learning_rate": 3e-06,
+      "loss": -0.1207,
+      "policy/loss": -0.42119356989860535,
+      "step": 2543
+    },
+    {
+      "epoch": 1.3584,
+      "grad_norm": 0.33766257762908936,
+      "learning_rate": 3e-06,
+      "loss": 0.016,
+      "policy/loss": -0.08246435225009918,
+      "step": 2544
+    },
+    {
+      "completion_length": 207.0416717529297,
+      "epoch": 1.3589333333333333,
+      "grad_norm": 0.30079153180122375,
+      "learning_rate": 3e-06,
+      "loss": 0.0046,
+      "policy/loss": -0.14804691076278687,
+      "reward": 1.0156250298023224,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6197916567325592,
+      "rewards/correctness_reward_func_math": 0.3958333283662796,
+      "step": 2545
+    },
+    {
+      "epoch": 1.3594666666666666,
+      "grad_norm": 0.31891006231307983,
+      "learning_rate": 3e-06,
+      "loss": -0.0179,
+      "policy/loss": -0.09773944690823555,
+      "step": 2546
+    },
+    {
+      "epoch": 1.3599999999999999,
+      "grad_norm": 0.28179121017456055,
+      "learning_rate": 3e-06,
+      "loss": 0.0031,
+      "policy/loss": -0.15567168593406677,
+      "step": 2547
+    },
+    {
+      "epoch": 1.3605333333333334,
+      "grad_norm": 0.31010544300079346,
+      "learning_rate": 3e-06,
+      "loss": -0.0196,
+      "policy/loss": -0.09843437932431698,
+      "step": 2548
+    },
+    {
+      "completion_length": 192.08334350585938,
+      "epoch": 1.3610666666666666,
+      "grad_norm": 0.23501494526863098,
+      "learning_rate": 3e-06,
+      "loss": -0.1719,
+      "policy/loss": -1.392854612566616,
+      "reward": 1.4531250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.7708333432674408,
+      "step": 2549
+    },
+    {
+      "epoch": 1.3616,
+      "grad_norm": 0.6664159297943115,
+      "learning_rate": 3e-06,
+      "loss": -0.1032,
+      "policy/loss": -0.4617436118118192,
+      "step": 2550
+    },
+    {
+      "epoch": 1.3621333333333334,
+      "grad_norm": 0.24135075509548187,
+      "learning_rate": 3e-06,
+      "loss": -0.1718,
+      "policy/loss": -1.3912976199559992,
+      "step": 2551
+    },
+    {
+      "epoch": 1.3626666666666667,
+      "grad_norm": 0.43949246406555176,
+      "learning_rate": 3e-06,
+      "loss": -0.1059,
+      "policy/loss": -0.47665305807485225,
+      "step": 2552
+    },
+    {
+      "completion_length": 204.77083587646484,
+      "epoch": 1.3632,
+      "grad_norm": 0.29985255002975464,
+      "learning_rate": 3e-06,
+      "loss": 0.0118,
+      "policy/loss": -0.1963988857920711,
+      "reward": 1.4322917461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 2553
+    },
+    {
+      "epoch": 1.3637333333333332,
+      "grad_norm": 0.2925895154476166,
+      "learning_rate": 3e-06,
+      "loss": 0.1666,
+      "policy/loss": 0.33921255533976336,
+      "step": 2554
+    },
+    {
+      "epoch": 1.3642666666666667,
+      "grad_norm": 0.3040183484554291,
+      "learning_rate": 3e-06,
+      "loss": 0.0104,
+      "policy/loss": -0.19541537935744557,
+      "step": 2555
+    },
+    {
+      "epoch": 1.3648,
+      "grad_norm": 0.29142192006111145,
+      "learning_rate": 3e-06,
+      "loss": 0.1659,
+      "policy/loss": 0.340004015643256,
+      "step": 2556
+    },
+    {
+      "completion_length": 204.40625762939453,
+      "epoch": 1.3653333333333333,
+      "grad_norm": 0.29566508531570435,
+      "learning_rate": 3e-06,
+      "loss": -0.1264,
+      "policy/loss": -0.04725807575165675,
+      "reward": 1.2552083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.5416666716337204,
+      "step": 2557
+    },
+    {
+      "epoch": 1.3658666666666668,
+      "grad_norm": 0.23032057285308838,
+      "learning_rate": 3e-06,
+      "loss": 0.2548,
+      "policy/loss": -0.030427041335919114,
+      "step": 2558
+    },
+    {
+      "epoch": 1.3664,
+      "grad_norm": 0.2935025691986084,
+      "learning_rate": 3e-06,
+      "loss": -0.1271,
+      "policy/loss": -0.04667318666878373,
+      "step": 2559
+    },
+    {
+      "epoch": 1.3669333333333333,
+      "grad_norm": 0.23914776742458344,
+      "learning_rate": 3e-06,
+      "loss": 0.254,
+      "policy/loss": -0.03162682333582012,
+      "step": 2560
+    },
+    {
+      "completion_length": 203.5729217529297,
+      "epoch": 1.3674666666666666,
+      "grad_norm": 0.30102840065956116,
+      "learning_rate": 3e-06,
+      "loss": 0.1218,
+      "policy/loss": -0.21092243497952978,
+      "reward": 1.244791716337204,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.5624999850988388,
+      "step": 2561
+    },
+    {
+      "epoch": 1.3679999999999999,
+      "grad_norm": 0.2569628655910492,
+      "learning_rate": 3e-06,
+      "loss": 0.0767,
+      "policy/loss": 0.28283149987937506,
+      "step": 2562
+    },
+    {
+      "epoch": 1.3685333333333334,
+      "grad_norm": 0.3038918972015381,
+      "learning_rate": 3e-06,
+      "loss": 0.1198,
+      "policy/loss": -0.2132975757828639,
+      "step": 2563
+    },
+    {
+      "epoch": 1.3690666666666667,
+      "grad_norm": 0.2298295497894287,
+      "learning_rate": 3e-06,
+      "loss": 0.0766,
+      "policy/loss": 0.2813399829198069,
+      "step": 2564
+    },
+    {
+      "completion_length": 217.96875,
+      "epoch": 1.3696,
+      "grad_norm": 0.29636532068252563,
+      "learning_rate": 3e-06,
+      "loss": 0.2218,
+      "policy/loss": 0.04497465747090956,
+      "reward": 1.5833333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333134651184,
+      "rewards/correctness_reward_func_math": 0.8750000298023224,
+      "step": 2565
+    },
+    {
+      "epoch": 1.3701333333333334,
+      "grad_norm": 0.4221133887767792,
+      "learning_rate": 3e-06,
+      "loss": 0.1387,
+      "policy/loss": 0.3801193825298945,
+      "step": 2566
+    },
+    {
+      "epoch": 1.3706666666666667,
+      "grad_norm": 0.2857709228992462,
+      "learning_rate": 3e-06,
+      "loss": 0.2209,
+      "policy/loss": 0.04561357844664826,
+      "step": 2567
+    },
+    {
+      "epoch": 1.3712,
+      "grad_norm": 0.2873064875602722,
+      "learning_rate": 3e-06,
+      "loss": 0.1361,
+      "policy/loss": 0.37983444258363264,
+      "step": 2568
+    },
+    {
+      "completion_length": 200.08333587646484,
+      "epoch": 1.3717333333333332,
+      "grad_norm": 0.22194981575012207,
+      "learning_rate": 3e-06,
+      "loss": 0.029,
+      "policy/loss": 0.12741558253765106,
+      "reward": 1.5989583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.8541666567325592,
+      "step": 2569
+    },
+    {
+      "epoch": 1.3722666666666667,
+      "grad_norm": 0.24620240926742554,
+      "learning_rate": 3e-06,
+      "loss": 0.0809,
+      "policy/loss": 0.17142640054225922,
+      "step": 2570
+    },
+    {
+      "epoch": 1.3728,
+      "grad_norm": 0.21777868270874023,
+      "learning_rate": 3e-06,
+      "loss": 0.0284,
+      "policy/loss": 0.1268995851278305,
+      "step": 2571
+    },
+    {
+      "epoch": 1.3733333333333333,
+      "grad_norm": 0.22165928781032562,
+      "learning_rate": 3e-06,
+      "loss": 0.0794,
+      "policy/loss": 0.17131510376930237,
+      "step": 2572
+    },
+    {
+      "completion_length": 201.30208587646484,
+      "epoch": 1.3738666666666668,
+      "grad_norm": 0.415969580411911,
+      "learning_rate": 3e-06,
+      "loss": -0.0596,
+      "policy/loss": 0.07462511729941923,
+      "reward": 1.3906250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.7083333432674408,
+      "step": 2573
+    },
+    {
+      "epoch": 1.3744,
+      "grad_norm": 0.2504129707813263,
+      "learning_rate": 3e-06,
+      "loss": -0.0351,
+      "policy/loss": 0.10798836100368447,
+      "step": 2574
+    },
+    {
+      "epoch": 1.3749333333333333,
+      "grad_norm": 0.3338586986064911,
+      "learning_rate": 3e-06,
+      "loss": -0.0605,
+      "policy/loss": 0.06853633110565482,
+      "step": 2575
+    },
+    {
+      "epoch": 1.3754666666666666,
+      "grad_norm": 0.2603647708892822,
+      "learning_rate": 3e-06,
+      "loss": -0.0364,
+      "policy/loss": 0.10590806222855775,
+      "step": 2576
+    },
+    {
+      "completion_length": 192.33334350585938,
+      "epoch": 1.376,
+      "grad_norm": 0.3161126673221588,
+      "learning_rate": 3e-06,
+      "loss": -0.1007,
+      "policy/loss": 1.6872686003921444e-07,
+      "reward": 1.34375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7395833134651184,
+      "rewards/correctness_reward_func_math": 0.6041666716337204,
+      "step": 2577
+    },
+    {
+      "epoch": 1.3765333333333334,
+      "grad_norm": 0.3812812268733978,
+      "learning_rate": 3e-06,
+      "loss": -0.0494,
+      "policy/loss": 2.0699716429817272e-07,
+      "step": 2578
+    },
+    {
+      "epoch": 1.3770666666666667,
+      "grad_norm": 0.3121418058872223,
+      "learning_rate": 3e-06,
+      "loss": -0.1019,
+      "policy/loss": 1.6858981055634104e-07,
+      "step": 2579
+    },
+    {
+      "epoch": 1.3776,
+      "grad_norm": 0.39546728134155273,
+      "learning_rate": 3e-06,
+      "loss": -0.0517,
+      "policy/loss": 2.0658207944279638e-07,
+      "step": 2580
+    },
+    {
+      "completion_length": 225.11459350585938,
+      "epoch": 1.3781333333333334,
+      "grad_norm": 0.5212864875793457,
+      "learning_rate": 3e-06,
+      "loss": -0.1429,
+      "policy/loss": -1.1106104254722595,
+      "reward": 1.4843750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.770833358168602,
+      "step": 2581
+    },
+    {
+      "epoch": 1.3786666666666667,
+      "grad_norm": 0.365560382604599,
+      "learning_rate": 3e-06,
+      "loss": -0.2331,
+      "policy/loss": -0.42342258244752884,
+      "step": 2582
+    },
+    {
+      "epoch": 1.3792,
+      "grad_norm": 0.3638833463191986,
+      "learning_rate": 3e-06,
+      "loss": -0.1448,
+      "policy/loss": -1.1161747574806213,
+      "step": 2583
+    },
+    {
+      "epoch": 1.3797333333333333,
+      "grad_norm": 0.3353373110294342,
+      "learning_rate": 3e-06,
+      "loss": -0.2349,
+      "policy/loss": -0.4330887719988823,
+      "step": 2584
+    },
+    {
+      "completion_length": 221.48958587646484,
+      "epoch": 1.3802666666666665,
+      "grad_norm": 0.2271396815776825,
+      "learning_rate": 3e-06,
+      "loss": 0.0297,
+      "policy/loss": -0.23165979236364365,
+      "reward": 1.4427083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.7291666865348816,
+      "step": 2585
+    },
+    {
+      "epoch": 1.3808,
+      "grad_norm": 0.2921297252178192,
+      "learning_rate": 3e-06,
+      "loss": -0.0018,
+      "policy/loss": 0.1334208846092224,
+      "step": 2586
+    },
+    {
+      "epoch": 1.3813333333333333,
+      "grad_norm": 0.2209780067205429,
+      "learning_rate": 3e-06,
+      "loss": 0.0293,
+      "policy/loss": -0.22967623732984066,
+      "step": 2587
+    },
+    {
+      "epoch": 1.3818666666666668,
+      "grad_norm": 0.2892264723777771,
+      "learning_rate": 3e-06,
+      "loss": -0.0025,
+      "policy/loss": 0.13241276144981384,
+      "step": 2588
+    },
+    {
+      "completion_length": 214.33333587646484,
+      "epoch": 1.3824,
+      "grad_norm": 0.33521440625190735,
+      "learning_rate": 3e-06,
+      "loss": 0.0181,
+      "policy/loss": 0.24825157242458928,
+      "reward": 1.5989583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.9166666865348816,
+      "step": 2589
+    },
+    {
+      "epoch": 1.3829333333333333,
+      "grad_norm": 0.3067467510700226,
+      "learning_rate": 3e-06,
+      "loss": -0.0076,
+      "policy/loss": -0.0026979850811414963,
+      "step": 2590
+    },
+    {
+      "epoch": 1.3834666666666666,
+      "grad_norm": 0.3375716507434845,
+      "learning_rate": 3e-06,
+      "loss": 0.0167,
+      "policy/loss": 0.24720643478172732,
+      "step": 2591
+    },
+    {
+      "epoch": 1.384,
+      "grad_norm": 0.30134353041648865,
+      "learning_rate": 3e-06,
+      "loss": -0.0074,
+      "policy/loss": -0.000524832923030516,
+      "step": 2592
+    },
+    {
+      "completion_length": 214.67708587646484,
+      "epoch": 1.3845333333333334,
+      "grad_norm": 0.3211832046508789,
+      "learning_rate": 3e-06,
+      "loss": 0.0742,
+      "policy/loss": 0.25116701424121857,
+      "reward": 1.1979167461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.5416666865348816,
+      "step": 2593
+    },
+    {
+      "epoch": 1.3850666666666667,
+      "grad_norm": 0.29506343603134155,
+      "learning_rate": 3e-06,
+      "loss": 0.0853,
+      "policy/loss": -0.3874510480090976,
+      "step": 2594
+    },
+    {
+      "epoch": 1.3856,
+      "grad_norm": 0.3292456567287445,
+      "learning_rate": 3e-06,
+      "loss": 0.0734,
+      "policy/loss": 0.2482365071773529,
+      "step": 2595
+    },
+    {
+      "epoch": 1.3861333333333334,
+      "grad_norm": 0.29430606961250305,
+      "learning_rate": 3e-06,
+      "loss": 0.0843,
+      "policy/loss": -0.3908099252730608,
+      "step": 2596
+    },
+    {
+      "completion_length": 186.36459350585938,
+      "epoch": 1.3866666666666667,
+      "grad_norm": 0.27745577692985535,
+      "learning_rate": 3e-06,
+      "loss": 0.1243,
+      "policy/loss": 0.023849710822105408,
+      "reward": 1.8281250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 1.0833333134651184,
+      "step": 2597
+    },
+    {
+      "epoch": 1.3872,
+      "grad_norm": 0.2662663459777832,
+      "learning_rate": 3e-06,
+      "loss": 0.0537,
+      "policy/loss": 0.17938506603240967,
+      "step": 2598
+    },
+    {
+      "epoch": 1.3877333333333333,
+      "grad_norm": 0.2859345078468323,
+      "learning_rate": 3e-06,
+      "loss": 0.1235,
+      "policy/loss": 0.025778312236070633,
+      "step": 2599
+    },
+    {
+      "epoch": 1.3882666666666665,
+      "grad_norm": 0.259243905544281,
+      "learning_rate": 3e-06,
+      "loss": 0.0531,
+      "policy/loss": 0.17934081703424454,
+      "step": 2600
+    },
+    {
+      "completion_length": 199.8125,
+      "epoch": 1.3888,
+      "grad_norm": 0.4828794002532959,
+      "learning_rate": 3e-06,
+      "loss": -0.1617,
+      "policy/loss": 0.38459272330570826,
+      "reward": 1.5052083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.7916666865348816,
+      "step": 2601
+    },
+    {
+      "epoch": 1.3893333333333333,
+      "grad_norm": 0.3157840073108673,
+      "learning_rate": 3e-06,
+      "loss": -0.0845,
+      "policy/loss": 0.30622989078003116,
+      "step": 2602
+    },
+    {
+      "epoch": 1.3898666666666666,
+      "grad_norm": 0.3605407178401947,
+      "learning_rate": 3e-06,
+      "loss": -0.1633,
+      "policy/loss": 0.3838773482836402,
+      "step": 2603
+    },
+    {
+      "epoch": 1.3904,
+      "grad_norm": 0.306033194065094,
+      "learning_rate": 3e-06,
+      "loss": -0.086,
+      "policy/loss": 0.30680692330882664,
+      "step": 2604
+    },
+    {
+      "completion_length": 197.0729217529297,
+      "epoch": 1.3909333333333334,
+      "grad_norm": 0.23218154907226562,
+      "learning_rate": 3e-06,
+      "loss": 0.0778,
+      "policy/loss": 0.14042501664082607,
+      "reward": 1.2708333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6458333432674408,
+      "rewards/correctness_reward_func_math": 0.6250000149011612,
+      "step": 2605
+    },
+    {
+      "epoch": 1.3914666666666666,
+      "grad_norm": 0.26302897930145264,
+      "learning_rate": 3e-06,
+      "loss": 0.0737,
+      "policy/loss": 0.094907087191622,
+      "step": 2606
+    },
+    {
+      "epoch": 1.392,
+      "grad_norm": 0.255402147769928,
+      "learning_rate": 3e-06,
+      "loss": 0.0769,
+      "policy/loss": 0.14083123714375034,
+      "step": 2607
+    },
+    {
+      "epoch": 1.3925333333333334,
+      "grad_norm": 0.3009352385997772,
+      "learning_rate": 3e-06,
+      "loss": 0.0722,
+      "policy/loss": 0.09520506571888632,
+      "step": 2608
+    },
+    {
+      "completion_length": 207.86459350585938,
+      "epoch": 1.3930666666666667,
+      "grad_norm": 0.2834857702255249,
+      "learning_rate": 3e-06,
+      "loss": -0.2154,
+      "policy/loss": 1.6889139686782073e-07,
+      "reward": 1.9375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 1.1875,
+      "step": 2609
+    },
+    {
+      "epoch": 1.3936,
+      "grad_norm": 0.46358585357666016,
+      "learning_rate": 3e-06,
+      "loss": -0.1218,
+      "policy/loss": 1.6289862969642854e-07,
+      "step": 2610
+    },
+    {
+      "epoch": 1.3941333333333334,
+      "grad_norm": 0.28326669335365295,
+      "learning_rate": 3e-06,
+      "loss": -0.2172,
+      "policy/loss": 1.6889315190837806e-07,
+      "step": 2611
+    },
+    {
+      "epoch": 1.3946666666666667,
+      "grad_norm": 0.47291862964630127,
+      "learning_rate": 3e-06,
+      "loss": -0.123,
+      "policy/loss": 1.6298805149972395e-07,
+      "step": 2612
+    },
+    {
+      "completion_length": 200.4479217529297,
+      "epoch": 1.3952,
+      "grad_norm": 0.25887787342071533,
+      "learning_rate": 3e-06,
+      "loss": 0.0318,
+      "policy/loss": 3.392813852087784e-08,
+      "reward": 1.3541666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 2613
+    },
+    {
+      "epoch": 1.3957333333333333,
+      "grad_norm": 0.28106242418289185,
+      "learning_rate": 3e-06,
+      "loss": -0.1011,
+      "policy/loss": 2.2944966815430234e-08,
+      "step": 2614
+    },
+    {
+      "epoch": 1.3962666666666665,
+      "grad_norm": 0.27020466327667236,
+      "learning_rate": 3e-06,
+      "loss": 0.0307,
+      "policy/loss": 3.3963537759973406e-08,
+      "step": 2615
+    },
+    {
+      "epoch": 1.3968,
+      "grad_norm": 0.29451456665992737,
+      "learning_rate": 3e-06,
+      "loss": -0.1019,
+      "policy/loss": 2.2942911570567048e-08,
+      "step": 2616
+    },
+    {
+      "completion_length": 203.7916717529297,
+      "epoch": 1.3973333333333333,
+      "grad_norm": 0.2593297064304352,
+      "learning_rate": 3e-06,
+      "loss": -0.1043,
+      "policy/loss": 0.26000034772057923,
+      "reward": 1.46875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.59375,
+      "rewards/correctness_reward_func_math": 0.875,
+      "step": 2617
+    },
+    {
+      "epoch": 1.3978666666666666,
+      "grad_norm": 0.2796626389026642,
+      "learning_rate": 3e-06,
+      "loss": 0.0045,
+      "policy/loss": 0.4750551888449621,
+      "step": 2618
+    },
+    {
+      "epoch": 1.3984,
+      "grad_norm": 0.26896393299102783,
+      "learning_rate": 3e-06,
+      "loss": -0.1053,
+      "policy/loss": 0.2580043372366596,
+      "step": 2619
+    },
+    {
+      "epoch": 1.3989333333333334,
+      "grad_norm": 0.265594482421875,
+      "learning_rate": 3e-06,
+      "loss": 0.0037,
+      "policy/loss": 0.4743130811821956,
+      "step": 2620
+    },
+    {
+      "completion_length": 206.9166717529297,
+      "epoch": 1.3994666666666666,
+      "grad_norm": 0.27222681045532227,
+      "learning_rate": 3e-06,
+      "loss": 0.1308,
+      "policy/loss": 1.3931264675193233e-07,
+      "reward": 1.3333333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 2621
+    },
+    {
+      "epoch": 1.4,
+      "grad_norm": 0.2571879029273987,
+      "learning_rate": 3e-06,
+      "loss": -0.0184,
+      "policy/loss": 1.5358963167955153e-07,
+      "step": 2622
+    },
+    {
+      "epoch": 1.4005333333333334,
+      "grad_norm": 0.2776772081851959,
+      "learning_rate": 3e-06,
+      "loss": 0.1295,
+      "policy/loss": 1.392580202264071e-07,
+      "step": 2623
+    },
+    {
+      "epoch": 1.4010666666666667,
+      "grad_norm": 0.26218318939208984,
+      "learning_rate": 3e-06,
+      "loss": -0.019,
+      "policy/loss": 1.5358884297711484e-07,
+      "step": 2624
+    },
+    {
+      "completion_length": 199.40625762939453,
+      "epoch": 1.4016,
+      "grad_norm": 0.28405290842056274,
+      "learning_rate": 3e-06,
+      "loss": -0.1817,
+      "policy/loss": -0.25125572830438614,
+      "reward": 1.6927083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.9791666567325592,
+      "step": 2625
+    },
+    {
+      "epoch": 1.4021333333333335,
+      "grad_norm": 0.26826003193855286,
+      "learning_rate": 3e-06,
+      "loss": -0.0674,
+      "policy/loss": 0.06086905440315604,
+      "step": 2626
+    },
+    {
+      "epoch": 1.4026666666666667,
+      "grad_norm": 0.26057687401771545,
+      "learning_rate": 3e-06,
+      "loss": -0.1834,
+      "policy/loss": -0.2584789991378784,
+      "step": 2627
+    },
+    {
+      "epoch": 1.4032,
+      "grad_norm": 0.27183273434638977,
+      "learning_rate": 3e-06,
+      "loss": -0.068,
+      "policy/loss": 0.059717971831560135,
+      "step": 2628
+    },
+    {
+      "completion_length": 208.95833587646484,
+      "epoch": 1.4037333333333333,
+      "grad_norm": 0.3615208864212036,
+      "learning_rate": 3e-06,
+      "loss": 0.0355,
+      "policy/loss": 0.379265982657671,
+      "reward": 1.2291666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6458333134651184,
+      "rewards/correctness_reward_func_math": 0.5833333283662796,
+      "step": 2629
+    },
+    {
+      "epoch": 1.4042666666666666,
+      "grad_norm": 0.32679036259651184,
+      "learning_rate": 3e-06,
+      "loss": 0.1756,
+      "policy/loss": 0.5897825807332993,
+      "step": 2630
+    },
+    {
+      "epoch": 1.4048,
+      "grad_norm": 0.2933906316757202,
+      "learning_rate": 3e-06,
+      "loss": 0.0345,
+      "policy/loss": 0.37753046676516533,
+      "step": 2631
+    },
+    {
+      "epoch": 1.4053333333333333,
+      "grad_norm": 0.29779088497161865,
+      "learning_rate": 3e-06,
+      "loss": 0.1743,
+      "policy/loss": 0.5847107768058777,
+      "step": 2632
+    },
+    {
+      "completion_length": 202.1666717529297,
+      "epoch": 1.4058666666666666,
+      "grad_norm": 0.2960178852081299,
+      "learning_rate": 3e-06,
+      "loss": -0.2467,
+      "policy/loss": -0.2580229267477989,
+      "reward": 1.2916667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7083333432674408,
+      "rewards/correctness_reward_func_math": 0.5833333432674408,
+      "step": 2633
+    },
+    {
+      "epoch": 1.4064,
+      "grad_norm": 0.2479606568813324,
+      "learning_rate": 3e-06,
+      "loss": -0.2425,
+      "policy/loss": -0.12951164692640305,
+      "step": 2634
+    },
+    {
+      "epoch": 1.4069333333333334,
+      "grad_norm": 0.3092445135116577,
+      "learning_rate": 3e-06,
+      "loss": -0.2488,
+      "policy/loss": -0.2600009962916374,
+      "step": 2635
+    },
+    {
+      "epoch": 1.4074666666666666,
+      "grad_norm": 0.24346165359020233,
+      "learning_rate": 3e-06,
+      "loss": -0.2439,
+      "policy/loss": -0.1317328829318285,
+      "step": 2636
+    },
+    {
+      "completion_length": 195.20834350585938,
+      "epoch": 1.408,
+      "grad_norm": 0.34590277075767517,
+      "learning_rate": 3e-06,
+      "loss": -0.0687,
+      "policy/loss": 7.426933379406364e-08,
+      "reward": 1.4114583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.640625,
+      "rewards/correctness_reward_func_math": 0.7708333432674408,
+      "step": 2637
+    },
+    {
+      "epoch": 1.4085333333333334,
+      "grad_norm": 0.2903572618961334,
+      "learning_rate": 3e-06,
+      "loss": 0.1618,
+      "policy/loss": 6.877744596067714e-08,
+      "step": 2638
+    },
+    {
+      "epoch": 1.4090666666666667,
+      "grad_norm": 0.3434552252292633,
+      "learning_rate": 3e-06,
+      "loss": -0.0698,
+      "policy/loss": 7.426281811717672e-08,
+      "step": 2639
+    },
+    {
+      "epoch": 1.4096,
+      "grad_norm": 0.3008776903152466,
+      "learning_rate": 3e-06,
+      "loss": 0.1606,
+      "policy/loss": 6.890891235400431e-08,
+      "step": 2640
+    },
+    {
+      "completion_length": 209.4166717529297,
+      "epoch": 1.4101333333333335,
+      "grad_norm": 0.2160850614309311,
+      "learning_rate": 3e-06,
+      "loss": 0.0755,
+      "policy/loss": 0.05764448925685173,
+      "reward": 1.6250000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.8749999701976776,
+      "step": 2641
+    },
+    {
+      "epoch": 1.4106666666666667,
+      "grad_norm": 0.2968295216560364,
+      "learning_rate": 3e-06,
+      "loss": -0.0482,
+      "policy/loss": -0.015996747384630083,
+      "step": 2642
+    },
+    {
+      "epoch": 1.4112,
+      "grad_norm": 0.21635864675045013,
+      "learning_rate": 3e-06,
+      "loss": 0.0736,
+      "policy/loss": 0.05372622894646639,
+      "step": 2643
+    },
+    {
+      "epoch": 1.4117333333333333,
+      "grad_norm": 0.2980264723300934,
+      "learning_rate": 3e-06,
+      "loss": -0.0503,
+      "policy/loss": -0.018948559351930072,
+      "step": 2644
+    },
+    {
+      "completion_length": 184.20833587646484,
+      "epoch": 1.4122666666666666,
+      "grad_norm": 0.2874331772327423,
+      "learning_rate": 3e-06,
+      "loss": 0.1258,
+      "policy/loss": 1.0621953848193115,
+      "reward": 1.5260416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.671875,
+      "rewards/correctness_reward_func_math": 0.8541666567325592,
+      "step": 2645
+    },
+    {
+      "epoch": 1.4128,
+      "grad_norm": 0.24836118519306183,
+      "learning_rate": 3e-06,
+      "loss": -0.0056,
+      "policy/loss": 0.07482582116219305,
+      "step": 2646
+    },
+    {
+      "epoch": 1.4133333333333333,
+      "grad_norm": 0.27681535482406616,
+      "learning_rate": 3e-06,
+      "loss": 0.1232,
+      "policy/loss": 1.0522585752650429,
+      "step": 2647
+    },
+    {
+      "epoch": 1.4138666666666666,
+      "grad_norm": 0.24151474237442017,
+      "learning_rate": 3e-06,
+      "loss": -0.0063,
+      "policy/loss": 0.07580207827560059,
+      "step": 2648
+    },
+    {
+      "completion_length": 206.1875,
+      "epoch": 1.4144,
+      "grad_norm": 0.5333224534988403,
+      "learning_rate": 3e-06,
+      "loss": 0.0905,
+      "policy/loss": 0.28301836799798963,
+      "reward": 1.3489583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6197916567325592,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 2649
+    },
+    {
+      "epoch": 1.4149333333333334,
+      "grad_norm": 0.29928961396217346,
+      "learning_rate": 3e-06,
+      "loss": 0.1348,
+      "policy/loss": -0.13849277371365787,
+      "step": 2650
+    },
+    {
+      "epoch": 1.4154666666666667,
+      "grad_norm": 0.3988916575908661,
+      "learning_rate": 3e-06,
+      "loss": 0.0891,
+      "policy/loss": 0.280884879514943,
+      "step": 2651
+    },
+    {
+      "epoch": 1.416,
+      "grad_norm": 0.29801061749458313,
+      "learning_rate": 3e-06,
+      "loss": 0.1344,
+      "policy/loss": -0.1391225710751769,
+      "step": 2652
+    },
+    {
+      "completion_length": 206.7291717529297,
+      "epoch": 1.4165333333333332,
+      "grad_norm": 0.4102952778339386,
+      "learning_rate": 3e-06,
+      "loss": 0.2395,
+      "policy/loss": 1.6175759753878083e-07,
+      "reward": 1.5364583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 2653
+    },
+    {
+      "epoch": 1.4170666666666667,
+      "grad_norm": 0.4635237753391266,
+      "learning_rate": 3e-06,
+      "loss": 0.0048,
+      "policy/loss": 1.3099871409849584e-07,
+      "step": 2654
+    },
+    {
+      "epoch": 1.4176,
+      "grad_norm": 0.3491719961166382,
+      "learning_rate": 3e-06,
+      "loss": 0.2386,
+      "policy/loss": 1.617555085431377e-07,
+      "step": 2655
+    },
+    {
+      "epoch": 1.4181333333333335,
+      "grad_norm": 0.31583550572395325,
+      "learning_rate": 3e-06,
+      "loss": 0.0044,
+      "policy/loss": 1.3103088392085738e-07,
+      "step": 2656
+    },
+    {
+      "completion_length": 192.7291717529297,
+      "epoch": 1.4186666666666667,
+      "grad_norm": 0.26501113176345825,
+      "learning_rate": 3e-06,
+      "loss": -0.2519,
+      "policy/loss": -0.7235957086086273,
+      "reward": 1.7812500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 1.0624999701976776,
+      "step": 2657
+    },
+    {
+      "epoch": 1.4192,
+      "grad_norm": 0.24747978150844574,
+      "learning_rate": 3e-06,
+      "loss": -0.1086,
+      "policy/loss": -0.02374391257762909,
+      "step": 2658
+    },
+    {
+      "epoch": 1.4197333333333333,
+      "grad_norm": 0.24212642014026642,
+      "learning_rate": 3e-06,
+      "loss": -0.2542,
+      "policy/loss": -0.7290602028369904,
+      "step": 2659
+    },
+    {
+      "epoch": 1.4202666666666666,
+      "grad_norm": 0.2504037320613861,
+      "learning_rate": 3e-06,
+      "loss": -0.1097,
+      "policy/loss": -0.023791514337062836,
+      "step": 2660
+    },
+    {
+      "completion_length": 197.12500762939453,
+      "epoch": 1.4208,
+      "grad_norm": 0.3403153419494629,
+      "learning_rate": 3e-06,
+      "loss": -0.0104,
+      "policy/loss": 4.275107023943292e-08,
+      "reward": 1.5885416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.875,
+      "step": 2661
+    },
+    {
+      "epoch": 1.4213333333333333,
+      "grad_norm": 0.3222307562828064,
+      "learning_rate": 3e-06,
+      "loss": 0.2057,
+      "policy/loss": 5.968712279980082e-08,
+      "step": 2662
+    },
+    {
+      "epoch": 1.4218666666666666,
+      "grad_norm": 0.33168405294418335,
+      "learning_rate": 3e-06,
+      "loss": -0.0112,
+      "policy/loss": 4.282553867085426e-08,
+      "step": 2663
+    },
+    {
+      "epoch": 1.4224,
+      "grad_norm": 0.33908843994140625,
+      "learning_rate": 3e-06,
+      "loss": 0.2054,
+      "policy/loss": 5.965361360438237e-08,
+      "step": 2664
+    },
+    {
+      "completion_length": 218.02083587646484,
+      "epoch": 1.4229333333333334,
+      "grad_norm": 0.2325381338596344,
+      "learning_rate": 3e-06,
+      "loss": -0.0517,
+      "policy/loss": -0.2856585130935869,
+      "reward": 1.010416716337204,
+      "rewards/boxed_and_answer_tags_format_reward": 0.59375,
+      "rewards/correctness_reward_func_math": 0.4166666716337204,
+      "step": 2665
+    },
+    {
+      "epoch": 1.4234666666666667,
+      "grad_norm": 0.2192109376192093,
+      "learning_rate": 3e-06,
+      "loss": -0.0729,
+      "policy/loss": -0.7793902509536821,
+      "step": 2666
+    },
+    {
+      "epoch": 1.424,
+      "grad_norm": 0.23611488938331604,
+      "learning_rate": 3e-06,
+      "loss": -0.0521,
+      "policy/loss": -0.2902677701357774,
+      "step": 2667
+    },
+    {
+      "epoch": 1.4245333333333332,
+      "grad_norm": 0.20202550292015076,
+      "learning_rate": 3e-06,
+      "loss": -0.0746,
+      "policy/loss": -0.7805872910942355,
+      "step": 2668
+    },
+    {
+      "completion_length": 216.4166717529297,
+      "epoch": 1.4250666666666667,
+      "grad_norm": 0.2454204261302948,
+      "learning_rate": 3e-06,
+      "loss": 0.075,
+      "policy/loss": 1.831273266361677e-07,
+      "reward": 1.4010416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 2669
+    },
+    {
+      "epoch": 1.4256,
+      "grad_norm": 0.5134155750274658,
+      "learning_rate": 3e-06,
+      "loss": -0.0747,
+      "policy/loss": 1.7270578922534696e-07,
+      "step": 2670
+    },
+    {
+      "epoch": 1.4261333333333333,
+      "grad_norm": 0.2656722366809845,
+      "learning_rate": 3e-06,
+      "loss": 0.0753,
+      "policy/loss": 1.831889306913581e-07,
+      "step": 2671
+    },
+    {
+      "epoch": 1.4266666666666667,
+      "grad_norm": 0.26102414727211,
+      "learning_rate": 3e-06,
+      "loss": -0.0745,
+      "policy/loss": 1.7268602192643812e-07,
+      "step": 2672
+    },
+    {
+      "completion_length": 209.2916717529297,
+      "epoch": 1.4272,
+      "grad_norm": 0.25718942284584045,
+      "learning_rate": 3e-06,
+      "loss": 0.1075,
+      "policy/loss": 8.691499076007858e-08,
+      "reward": 1.28125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 2673
+    },
+    {
+      "epoch": 1.4277333333333333,
+      "grad_norm": 0.25146713852882385,
+      "learning_rate": 3e-06,
+      "loss": -0.0403,
+      "policy/loss": 7.45817931857573e-08,
+      "step": 2674
+    },
+    {
+      "epoch": 1.4282666666666666,
+      "grad_norm": 0.2569669187068939,
+      "learning_rate": 3e-06,
+      "loss": 0.106,
+      "policy/loss": 8.698869891077265e-08,
+      "step": 2675
+    },
+    {
+      "epoch": 1.4288,
+      "grad_norm": 0.2508297264575958,
+      "learning_rate": 3e-06,
+      "loss": -0.0424,
+      "policy/loss": 7.454788253369315e-08,
+      "step": 2676
+    },
+    {
+      "completion_length": 217.11459350585938,
+      "epoch": 1.4293333333333333,
+      "grad_norm": 0.3896558880805969,
+      "learning_rate": 3e-06,
+      "loss": -0.0048,
+      "policy/loss": 1.8688374581188327e-07,
+      "reward": 1.6510416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.9375,
+      "step": 2677
+    },
+    {
+      "epoch": 1.4298666666666666,
+      "grad_norm": 0.26455777883529663,
+      "learning_rate": 3e-06,
+      "loss": -0.0977,
+      "policy/loss": 2.182055069965827e-07,
+      "step": 2678
+    },
+    {
+      "epoch": 1.4304000000000001,
+      "grad_norm": 0.37981921434402466,
+      "learning_rate": 3e-06,
+      "loss": -0.0063,
+      "policy/loss": 1.868580667974129e-07,
+      "step": 2679
+    },
+    {
+      "epoch": 1.4309333333333334,
+      "grad_norm": 0.2751743495464325,
+      "learning_rate": 3e-06,
+      "loss": -0.0988,
+      "policy/loss": 2.1801432126267173e-07,
+      "step": 2680
+    },
+    {
+      "completion_length": 214.86458587646484,
+      "epoch": 1.4314666666666667,
+      "grad_norm": 0.2917224168777466,
+      "learning_rate": 3e-06,
+      "loss": -0.0135,
+      "policy/loss": -0.30737993411976383,
+      "reward": 1.7864583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 1.0416666567325592,
+      "step": 2681
+    },
+    {
+      "epoch": 1.432,
+      "grad_norm": 0.4365779459476471,
+      "learning_rate": 3e-06,
+      "loss": 0.1968,
+      "policy/loss": 0.6687126515684625,
+      "step": 2682
+    },
+    {
+      "epoch": 1.4325333333333332,
+      "grad_norm": 0.2855347990989685,
+      "learning_rate": 3e-06,
+      "loss": -0.0141,
+      "policy/loss": -0.3123816280303018,
+      "step": 2683
+    },
+    {
+      "epoch": 1.4330666666666667,
+      "grad_norm": 0.4507347345352173,
+      "learning_rate": 3e-06,
+      "loss": 0.1944,
+      "policy/loss": 0.6665652749984083,
+      "step": 2684
+    },
+    {
+      "completion_length": 220.39583587646484,
+      "epoch": 1.4336,
+      "grad_norm": 0.30335453152656555,
+      "learning_rate": 3e-06,
+      "loss": 0.0683,
+      "policy/loss": 0.6321659684181213,
+      "reward": 1.2916666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.5416666567325592,
+      "step": 2685
+    },
+    {
+      "epoch": 1.4341333333333333,
+      "grad_norm": 0.27076414227485657,
+      "learning_rate": 3e-06,
+      "loss": -0.0329,
+      "policy/loss": 0.4389841528609395,
+      "step": 2686
+    },
+    {
+      "epoch": 1.4346666666666668,
+      "grad_norm": 0.2818685472011566,
+      "learning_rate": 3e-06,
+      "loss": 0.0683,
+      "policy/loss": 0.6251501739025116,
+      "step": 2687
+    },
+    {
+      "epoch": 1.4352,
+      "grad_norm": 0.26676037907600403,
+      "learning_rate": 3e-06,
+      "loss": -0.0329,
+      "policy/loss": 0.44043151941150427,
+      "step": 2688
+    },
+    {
+      "completion_length": 220.52083587646484,
+      "epoch": 1.4357333333333333,
+      "grad_norm": 0.24140523374080658,
+      "learning_rate": 3e-06,
+      "loss": -0.1369,
+      "policy/loss": 7.862328210705982e-08,
+      "reward": 1.1822917461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.5000000149011612,
+      "step": 2689
+    },
+    {
+      "epoch": 1.4362666666666666,
+      "grad_norm": 0.2577459514141083,
+      "learning_rate": 3e-06,
+      "loss": -0.1785,
+      "policy/loss": 9.535927958381762e-08,
+      "step": 2690
+    },
+    {
+      "epoch": 1.4368,
+      "grad_norm": 0.25277969241142273,
+      "learning_rate": 3e-06,
+      "loss": -0.1379,
+      "policy/loss": 7.871945939541547e-08,
+      "step": 2691
+    },
+    {
+      "epoch": 1.4373333333333334,
+      "grad_norm": 0.2805227041244507,
+      "learning_rate": 3e-06,
+      "loss": -0.1799,
+      "policy/loss": 9.543043510973348e-08,
+      "step": 2692
+    },
+    {
+      "completion_length": 209.21875762939453,
+      "epoch": 1.4378666666666666,
+      "grad_norm": 0.25919875502586365,
+      "learning_rate": 3e-06,
+      "loss": 0.2229,
+      "policy/loss": 0.09836555098060273,
+      "reward": 1.5260416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.875,
+      "step": 2693
+    },
+    {
+      "epoch": 1.4384000000000001,
+      "grad_norm": 0.32860758900642395,
+      "learning_rate": 3e-06,
+      "loss": 0.1095,
+      "policy/loss": 0.06281600505397478,
+      "step": 2694
+    },
+    {
+      "epoch": 1.4389333333333334,
+      "grad_norm": 0.25820299983024597,
+      "learning_rate": 3e-06,
+      "loss": 0.2229,
+      "policy/loss": 0.09895311125468709,
+      "step": 2695
+    },
+    {
+      "epoch": 1.4394666666666667,
+      "grad_norm": 0.3323052227497101,
+      "learning_rate": 3e-06,
+      "loss": 0.1094,
+      "policy/loss": 0.06268274408635932,
+      "step": 2696
+    },
+    {
+      "completion_length": 212.48959350585938,
+      "epoch": 1.44,
+      "grad_norm": 0.2367558777332306,
+      "learning_rate": 3e-06,
+      "loss": 0.0689,
+      "policy/loss": 1.5328127034308636e-07,
+      "reward": 1.4687500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 2697
+    },
+    {
+      "epoch": 1.4405333333333332,
+      "grad_norm": 0.25948119163513184,
+      "learning_rate": 3e-06,
+      "loss": 0.0247,
+      "policy/loss": 1.6566053773203748e-07,
+      "step": 2698
+    },
+    {
+      "epoch": 1.4410666666666667,
+      "grad_norm": 0.25259488821029663,
+      "learning_rate": 3e-06,
+      "loss": 0.0671,
+      "policy/loss": 1.5338771675033058e-07,
+      "step": 2699
+    },
+    {
+      "epoch": 1.4416,
+      "grad_norm": 0.2563762664794922,
+      "learning_rate": 3e-06,
+      "loss": 0.0245,
+      "policy/loss": 1.6576294115111523e-07,
+      "step": 2700
+    },
+    {
+      "completion_length": 218.8229217529297,
+      "epoch": 1.4421333333333333,
+      "grad_norm": 0.21063190698623657,
+      "learning_rate": 3e-06,
+      "loss": -0.1646,
+      "policy/loss": 0.040287289026202266,
+      "reward": 1.2187500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.5624999850988388,
+      "step": 2701
+    },
+    {
+      "epoch": 1.4426666666666668,
+      "grad_norm": 0.2685355544090271,
+      "learning_rate": 3e-06,
+      "loss": -0.1246,
+      "policy/loss": -0.3100661791112236,
+      "step": 2702
+    },
+    {
+      "epoch": 1.4432,
+      "grad_norm": 0.218924880027771,
+      "learning_rate": 3e-06,
+      "loss": -0.1664,
+      "policy/loss": 0.035149640600311116,
+      "step": 2703
+    },
+    {
+      "epoch": 1.4437333333333333,
+      "grad_norm": 0.2767060697078705,
+      "learning_rate": 3e-06,
+      "loss": -0.1259,
+      "policy/loss": -0.31097196109269376,
+      "step": 2704
+    },
+    {
+      "completion_length": 227.58334350585938,
+      "epoch": 1.4442666666666666,
+      "grad_norm": 0.3062973618507385,
+      "learning_rate": 3e-06,
+      "loss": 0.0603,
+      "policy/loss": 1.4101414791412026e-07,
+      "reward": 0.875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.25,
+      "step": 2705
+    },
+    {
+      "epoch": 1.4447999999999999,
+      "grad_norm": 0.28011777997016907,
+      "learning_rate": 3e-06,
+      "loss": -0.0248,
+      "policy/loss": 1.4552330185324536e-07,
+      "step": 2706
+    },
+    {
+      "epoch": 1.4453333333333334,
+      "grad_norm": 0.29235222935676575,
+      "learning_rate": 3e-06,
+      "loss": 0.0585,
+      "policy/loss": 1.4106802836977295e-07,
+      "step": 2707
+    },
+    {
+      "epoch": 1.4458666666666666,
+      "grad_norm": 0.2946344316005707,
+      "learning_rate": 3e-06,
+      "loss": -0.0252,
+      "policy/loss": 1.4553754468238367e-07,
+      "step": 2708
+    },
+    {
+      "completion_length": 230.58334350585938,
+      "epoch": 1.4464000000000001,
+      "grad_norm": 0.27198758721351624,
+      "learning_rate": 3e-06,
+      "loss": -0.1048,
+      "policy/loss": -0.2200364417939369,
+      "reward": 1.4166666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 2709
+    },
+    {
+      "epoch": 1.4469333333333334,
+      "grad_norm": 0.2688281834125519,
+      "learning_rate": 3e-06,
+      "loss": -0.0053,
+      "policy/loss": -0.2338150922058837,
+      "step": 2710
+    },
+    {
+      "epoch": 1.4474666666666667,
+      "grad_norm": 0.2675100266933441,
+      "learning_rate": 3e-06,
+      "loss": -0.1048,
+      "policy/loss": -0.22075397746224823,
+      "step": 2711
+    },
+    {
+      "epoch": 1.448,
+      "grad_norm": 0.2611140012741089,
+      "learning_rate": 3e-06,
+      "loss": -0.007,
+      "policy/loss": -0.2345748577217659,
+      "step": 2712
+    },
+    {
+      "completion_length": 217.65625,
+      "epoch": 1.4485333333333332,
+      "grad_norm": 0.18120039999485016,
+      "learning_rate": 3e-06,
+      "loss": 0.0747,
+      "policy/loss": 5.829283633573823e-08,
+      "reward": 1.6927083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.9791666567325592,
+      "step": 2713
+    },
+    {
+      "epoch": 1.4490666666666667,
+      "grad_norm": 0.349947065114975,
+      "learning_rate": 3e-06,
+      "loss": 0.0052,
+      "policy/loss": 1.1533319721479529e-07,
+      "step": 2714
+    },
+    {
+      "epoch": 1.4496,
+      "grad_norm": 0.18336229026317596,
+      "learning_rate": 3e-06,
+      "loss": 0.074,
+      "policy/loss": 5.8141353953544694e-08,
+      "step": 2715
+    },
+    {
+      "epoch": 1.4501333333333333,
+      "grad_norm": 0.3341595232486725,
+      "learning_rate": 3e-06,
+      "loss": 0.0041,
+      "policy/loss": 1.1540079114524815e-07,
+      "step": 2716
+    },
+    {
+      "completion_length": 201.3229217529297,
+      "epoch": 1.4506666666666668,
+      "grad_norm": 0.27928978204727173,
+      "learning_rate": 3e-06,
+      "loss": 0.1199,
+      "policy/loss": 0.18161118290252887,
+      "reward": 1.7083333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 1.0208333432674408,
+      "step": 2717
+    },
+    {
+      "epoch": 1.4512,
+      "grad_norm": 0.2498231679201126,
+      "learning_rate": 3e-06,
+      "loss": -0.0722,
+      "policy/loss": -0.21743079141217514,
+      "step": 2718
+    },
+    {
+      "epoch": 1.4517333333333333,
+      "grad_norm": 0.27191534638404846,
+      "learning_rate": 3e-06,
+      "loss": 0.1191,
+      "policy/loss": 0.1828261044394992,
+      "step": 2719
+    },
+    {
+      "epoch": 1.4522666666666666,
+      "grad_norm": 0.2676312327384949,
+      "learning_rate": 3e-06,
+      "loss": -0.0731,
+      "policy/loss": -0.21948984875344202,
+      "step": 2720
+    },
+    {
+      "completion_length": 214.40625,
+      "epoch": 1.4527999999999999,
+      "grad_norm": 0.2332872599363327,
+      "learning_rate": 3e-06,
+      "loss": 0.1011,
+      "policy/loss": -0.003046264514885877,
+      "reward": 1.6354166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.9791666567325592,
+      "step": 2721
+    },
+    {
+      "epoch": 1.4533333333333334,
+      "grad_norm": 0.3401041030883789,
+      "learning_rate": 3e-06,
+      "loss": -0.0447,
+      "policy/loss": 0.011578651562253128,
+      "step": 2722
+    },
+    {
+      "epoch": 1.4538666666666666,
+      "grad_norm": 0.24421343207359314,
+      "learning_rate": 3e-06,
+      "loss": 0.0998,
+      "policy/loss": -0.0026652300803817752,
+      "step": 2723
+    },
+    {
+      "epoch": 1.4544000000000001,
+      "grad_norm": 0.2995254695415497,
+      "learning_rate": 3e-06,
+      "loss": -0.045,
+      "policy/loss": 0.01196485319135121,
+      "step": 2724
+    },
+    {
+      "completion_length": 219.37500762939453,
+      "epoch": 1.4549333333333334,
+      "grad_norm": 0.27750682830810547,
+      "learning_rate": 3e-06,
+      "loss": 0.0005,
+      "policy/loss": -0.49615707993507385,
+      "reward": 1.5364583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.8541666865348816,
+      "step": 2725
+    },
+    {
+      "epoch": 1.4554666666666667,
+      "grad_norm": 0.30455881357192993,
+      "learning_rate": 3e-06,
+      "loss": 0.0126,
+      "policy/loss": -0.7035402655601501,
+      "step": 2726
+    },
+    {
+      "epoch": 1.456,
+      "grad_norm": 0.2865932881832123,
+      "learning_rate": 3e-06,
+      "loss": -0.0002,
+      "policy/loss": -0.49819666147232056,
+      "step": 2727
+    },
+    {
+      "epoch": 1.4565333333333332,
+      "grad_norm": 0.31004762649536133,
+      "learning_rate": 3e-06,
+      "loss": 0.0106,
+      "policy/loss": -0.7055199146270752,
+      "step": 2728
+    },
+    {
+      "completion_length": 213.5104217529297,
+      "epoch": 1.4570666666666667,
+      "grad_norm": 0.271331787109375,
+      "learning_rate": 3e-06,
+      "loss": 0.076,
+      "policy/loss": -0.49760377137274503,
+      "reward": 1.2395833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.5833333432674408,
+      "step": 2729
+    },
+    {
+      "epoch": 1.4576,
+      "grad_norm": 0.3312097489833832,
+      "learning_rate": 3e-06,
+      "loss": 0.0627,
+      "policy/loss": -0.028933950915678963,
+      "step": 2730
+    },
+    {
+      "epoch": 1.4581333333333333,
+      "grad_norm": 0.2572011649608612,
+      "learning_rate": 3e-06,
+      "loss": 0.0748,
+      "policy/loss": -0.4970236393832792,
+      "step": 2731
+    },
+    {
+      "epoch": 1.4586666666666668,
+      "grad_norm": 0.386488676071167,
+      "learning_rate": 3e-06,
+      "loss": 0.0613,
+      "policy/loss": -0.030020932820377766,
+      "step": 2732
+    },
+    {
+      "completion_length": 202.5729217529297,
+      "epoch": 1.4592,
+      "grad_norm": 0.2919173836708069,
+      "learning_rate": 3e-06,
+      "loss": 0.0419,
+      "policy/loss": -0.782102079600028,
+      "reward": 1.5052083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.7916666716337204,
+      "step": 2733
+    },
+    {
+      "epoch": 1.4597333333333333,
+      "grad_norm": 0.22569216787815094,
+      "learning_rate": 3e-06,
+      "loss": -0.0098,
+      "policy/loss": -0.9961935540443925,
+      "step": 2734
+    },
+    {
+      "epoch": 1.4602666666666666,
+      "grad_norm": 0.29645416140556335,
+      "learning_rate": 3e-06,
+      "loss": 0.0399,
+      "policy/loss": -0.7835644199522438,
+      "step": 2735
+    },
+    {
+      "epoch": 1.4607999999999999,
+      "grad_norm": 0.22830483317375183,
+      "learning_rate": 3e-06,
+      "loss": -0.0104,
+      "policy/loss": -0.9966900606915488,
+      "step": 2736
+    },
+    {
+      "completion_length": 220.96875762939453,
+      "epoch": 1.4613333333333334,
+      "grad_norm": 0.2724457085132599,
+      "learning_rate": 3e-06,
+      "loss": 0.1099,
+      "policy/loss": 0.17994399387494298,
+      "reward": 1.9375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 1.1875,
+      "step": 2737
+    },
+    {
+      "epoch": 1.4618666666666666,
+      "grad_norm": 0.25434765219688416,
+      "learning_rate": 3e-06,
+      "loss": 0.0212,
+      "policy/loss": -0.2091162056337481,
+      "step": 2738
+    },
+    {
+      "epoch": 1.4624,
+      "grad_norm": 0.28005874156951904,
+      "learning_rate": 3e-06,
+      "loss": 0.1086,
+      "policy/loss": 0.17793442324872188,
+      "step": 2739
+    },
+    {
+      "epoch": 1.4629333333333334,
+      "grad_norm": 0.3054256737232208,
+      "learning_rate": 3e-06,
+      "loss": 0.0201,
+      "policy/loss": -0.2075354011244599,
+      "step": 2740
+    },
+    {
+      "completion_length": 196.59375762939453,
+      "epoch": 1.4634666666666667,
+      "grad_norm": 0.42240259051322937,
+      "learning_rate": 3e-06,
+      "loss": -0.211,
+      "policy/loss": -0.2709603721316789,
+      "reward": 1.0312500298023224,
+      "rewards/boxed_and_answer_tags_format_reward": 0.53125,
+      "rewards/correctness_reward_func_math": 0.4999999850988388,
+      "step": 2741
+    },
+    {
+      "epoch": 1.464,
+      "grad_norm": 0.3301052153110504,
+      "learning_rate": 3e-06,
+      "loss": 0.1354,
+      "policy/loss": 0.5278943612922085,
+      "step": 2742
+    },
+    {
+      "epoch": 1.4645333333333332,
+      "grad_norm": 0.4224390685558319,
+      "learning_rate": 3e-06,
+      "loss": -0.213,
+      "policy/loss": -0.26975185818267633,
+      "step": 2743
+    },
+    {
+      "epoch": 1.4650666666666667,
+      "grad_norm": 0.3157171905040741,
+      "learning_rate": 3e-06,
+      "loss": 0.1341,
+      "policy/loss": 0.5265352561169081,
+      "step": 2744
+    },
+    {
+      "completion_length": 216.67709350585938,
+      "epoch": 1.4656,
+      "grad_norm": 0.3506055772304535,
+      "learning_rate": 3e-06,
+      "loss": -0.0999,
+      "policy/loss": 1.1976186087281349e-07,
+      "reward": 1.3333333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.6458333283662796,
+      "step": 2745
+    },
+    {
+      "epoch": 1.4661333333333333,
+      "grad_norm": 0.28884389996528625,
+      "learning_rate": 3e-06,
+      "loss": 0.1016,
+      "policy/loss": 1.0791209703597815e-07,
+      "step": 2746
+    },
+    {
+      "epoch": 1.4666666666666668,
+      "grad_norm": 0.31594136357307434,
+      "learning_rate": 3e-06,
+      "loss": -0.1012,
+      "policy/loss": 1.1968273838647292e-07,
+      "step": 2747
+    },
+    {
+      "epoch": 1.4672,
+      "grad_norm": 0.3125893175601959,
+      "learning_rate": 3e-06,
+      "loss": 0.0999,
+      "policy/loss": 1.0799616134704593e-07,
+      "step": 2748
+    },
+    {
+      "completion_length": 220.55208587646484,
+      "epoch": 1.4677333333333333,
+      "grad_norm": 0.2844066917896271,
+      "learning_rate": 3e-06,
+      "loss": 0.0876,
+      "policy/loss": 1.581529893712741e-07,
+      "reward": 1.3541667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.6666666716337204,
+      "step": 2749
+    },
+    {
+      "epoch": 1.4682666666666666,
+      "grad_norm": 0.34847816824913025,
+      "learning_rate": 3e-06,
+      "loss": 0.1476,
+      "policy/loss": 1.3691550648786688e-07,
+      "step": 2750
+    },
+    {
+      "epoch": 1.4687999999999999,
+      "grad_norm": 0.2986908257007599,
+      "learning_rate": 3e-06,
+      "loss": 0.0876,
+      "policy/loss": 1.5809978748393405e-07,
+      "step": 2751
+    },
+    {
+      "epoch": 1.4693333333333334,
+      "grad_norm": 0.35695064067840576,
+      "learning_rate": 3e-06,
+      "loss": 0.1463,
+      "policy/loss": 1.369877828949484e-07,
+      "step": 2752
+    },
+    {
+      "completion_length": 227.21875,
+      "epoch": 1.4698666666666667,
+      "grad_norm": 0.8266853094100952,
+      "learning_rate": 3e-06,
+      "loss": -0.1159,
+      "policy/loss": 8.586803801335918e-08,
+      "reward": 1.5208333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.7708333432674408,
+      "step": 2753
+    },
+    {
+      "epoch": 1.4704,
+      "grad_norm": 0.3303491473197937,
+      "learning_rate": 3e-06,
+      "loss": -0.0447,
+      "policy/loss": 9.485284024890461e-08,
+      "step": 2754
+    },
+    {
+      "epoch": 1.4709333333333334,
+      "grad_norm": 0.3196667432785034,
+      "learning_rate": 3e-06,
+      "loss": -0.1148,
+      "policy/loss": 8.588277111698517e-08,
+      "step": 2755
+    },
+    {
+      "epoch": 1.4714666666666667,
+      "grad_norm": 0.30832183361053467,
+      "learning_rate": 3e-06,
+      "loss": -0.0459,
+      "policy/loss": 9.522727850708179e-08,
+      "step": 2756
+    },
+    {
+      "completion_length": 227.86459350585938,
+      "epoch": 1.472,
+      "grad_norm": 0.23147054016590118,
+      "learning_rate": 3e-06,
+      "loss": -0.26,
+      "policy/loss": -0.31813571826224063,
+      "reward": 1.2500000596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.6250000149011612,
+      "step": 2757
+    },
+    {
+      "epoch": 1.4725333333333332,
+      "grad_norm": 0.29318276047706604,
+      "learning_rate": 3e-06,
+      "loss": 0.0036,
+      "policy/loss": -0.047578919633284755,
+      "step": 2758
+    },
+    {
+      "epoch": 1.4730666666666667,
+      "grad_norm": 0.23984572291374207,
+      "learning_rate": 3e-06,
+      "loss": -0.261,
+      "policy/loss": -0.3225726582273296,
+      "step": 2759
+    },
+    {
+      "epoch": 1.4736,
+      "grad_norm": 0.2807272672653198,
+      "learning_rate": 3e-06,
+      "loss": 0.0021,
+      "policy/loss": -0.05003703377968094,
+      "step": 2760
+    },
+    {
+      "completion_length": 225.0104217529297,
+      "epoch": 1.4741333333333333,
+      "grad_norm": 0.2853645980358124,
+      "learning_rate": 3e-06,
+      "loss": 0.0062,
+      "policy/loss": 1.2328047915843854e-07,
+      "reward": 1.1718750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6302083432674408,
+      "rewards/correctness_reward_func_math": 0.5416666716337204,
+      "step": 2761
+    },
+    {
+      "epoch": 1.4746666666666668,
+      "grad_norm": 0.30580195784568787,
+      "learning_rate": 3e-06,
+      "loss": -0.0263,
+      "policy/loss": 1.394715241076483e-07,
+      "step": 2762
+    },
+    {
+      "epoch": 1.4752,
+      "grad_norm": 0.2827370762825012,
+      "learning_rate": 3e-06,
+      "loss": 0.0046,
+      "policy/loss": 1.2330172083352409e-07,
+      "step": 2763
+    },
+    {
+      "epoch": 1.4757333333333333,
+      "grad_norm": 0.29713118076324463,
+      "learning_rate": 3e-06,
+      "loss": -0.0272,
+      "policy/loss": 1.3947768451316733e-07,
+      "step": 2764
+    },
+    {
+      "completion_length": 219.3541717529297,
+      "epoch": 1.4762666666666666,
+      "grad_norm": 0.28999513387680054,
+      "learning_rate": 3e-06,
+      "loss": 0.015,
+      "policy/loss": 1.3548331168067307e-07,
+      "reward": 1.3645833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.6458333432674408,
+      "step": 2765
+    },
+    {
+      "epoch": 1.4768,
+      "grad_norm": 0.4815675914287567,
+      "learning_rate": 3e-06,
+      "loss": -0.0958,
+      "policy/loss": 1.0994123300633873e-07,
+      "step": 2766
+    },
+    {
+      "epoch": 1.4773333333333334,
+      "grad_norm": 0.3103138506412506,
+      "learning_rate": 3e-06,
+      "loss": 0.0137,
+      "policy/loss": 1.352868466142354e-07,
+      "step": 2767
+    },
+    {
+      "epoch": 1.4778666666666667,
+      "grad_norm": 0.2911204993724823,
+      "learning_rate": 3e-06,
+      "loss": -0.0959,
+      "policy/loss": 1.0990902055141305e-07,
+      "step": 2768
+    },
+    {
+      "completion_length": 229.31250762939453,
+      "epoch": 1.4784,
+      "grad_norm": 0.2619846761226654,
+      "learning_rate": 3e-06,
+      "loss": 0.0105,
+      "policy/loss": 0.3853195570409298,
+      "reward": 1.5781250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.8333333432674408,
+      "step": 2769
+    },
+    {
+      "epoch": 1.4789333333333334,
+      "grad_norm": 0.38358718156814575,
+      "learning_rate": 3e-06,
+      "loss": 0.0272,
+      "policy/loss": 0.492265984416008,
+      "step": 2770
+    },
+    {
+      "epoch": 1.4794666666666667,
+      "grad_norm": 0.2636324167251587,
+      "learning_rate": 3e-06,
+      "loss": 0.0092,
+      "policy/loss": 0.3858572244644165,
+      "step": 2771
+    },
+    {
+      "epoch": 1.48,
+      "grad_norm": 0.4169829487800598,
+      "learning_rate": 3e-06,
+      "loss": 0.0268,
+      "policy/loss": 0.49279364943504333,
+      "step": 2772
+    },
+    {
+      "completion_length": 231.55209350585938,
+      "epoch": 1.4805333333333333,
+      "grad_norm": 0.30762580037117004,
+      "learning_rate": 3e-06,
+      "loss": 0.0487,
+      "policy/loss": -0.012605828698724508,
+      "reward": 1.2395833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6770833432674408,
+      "rewards/correctness_reward_func_math": 0.5625,
+      "step": 2773
+    },
+    {
+      "epoch": 1.4810666666666665,
+      "grad_norm": 0.35917264223098755,
+      "learning_rate": 3e-06,
+      "loss": 0.0003,
+      "policy/loss": 0.013391923159360886,
+      "step": 2774
+    },
+    {
+      "epoch": 1.4816,
+      "grad_norm": 0.30104920268058777,
+      "learning_rate": 3e-06,
+      "loss": 0.0466,
+      "policy/loss": -0.012208366766571999,
+      "step": 2775
+    },
+    {
+      "epoch": 1.4821333333333333,
+      "grad_norm": 0.3501822352409363,
+      "learning_rate": 3e-06,
+      "loss": -0.0013,
+      "policy/loss": 0.012134432792663574,
+      "step": 2776
+    },
+    {
+      "completion_length": 202.93750762939453,
+      "epoch": 1.4826666666666668,
+      "grad_norm": 0.4627557098865509,
+      "learning_rate": 3e-06,
+      "loss": 0.1022,
+      "policy/loss": 1.217472203052239e-07,
+      "reward": 1.4635416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 2777
+    },
+    {
+      "epoch": 1.4832,
+      "grad_norm": 0.2855207026004791,
+      "learning_rate": 3e-06,
+      "loss": -0.0594,
+      "policy/loss": 1.1311389158663587e-07,
+      "step": 2778
+    },
+    {
+      "epoch": 1.4837333333333333,
+      "grad_norm": 0.49620842933654785,
+      "learning_rate": 3e-06,
+      "loss": 0.0997,
+      "policy/loss": 1.2178153241393375e-07,
+      "step": 2779
+    },
+    {
+      "epoch": 1.4842666666666666,
+      "grad_norm": 0.45554864406585693,
+      "learning_rate": 3e-06,
+      "loss": -0.0603,
+      "policy/loss": 1.131204392379459e-07,
+      "step": 2780
+    },
+    {
+      "completion_length": 229.28125762939453,
+      "epoch": 1.4848,
+      "grad_norm": 0.2828729748725891,
+      "learning_rate": 3e-06,
+      "loss": -0.116,
+      "policy/loss": 1.1488248219393427e-07,
+      "reward": 1.40625,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 2781
+    },
+    {
+      "epoch": 1.4853333333333334,
+      "grad_norm": 0.2416212409734726,
+      "learning_rate": 3e-06,
+      "loss": -0.012,
+      "policy/loss": 1.42227992938615e-07,
+      "step": 2782
+    },
+    {
+      "epoch": 1.4858666666666667,
+      "grad_norm": 0.31529656052589417,
+      "learning_rate": 3e-06,
+      "loss": -0.1185,
+      "policy/loss": 1.1514283926317148e-07,
+      "step": 2783
+    },
+    {
+      "epoch": 1.4864,
+      "grad_norm": 0.3037480115890503,
+      "learning_rate": 3e-06,
+      "loss": -0.0131,
+      "policy/loss": 1.424385800419259e-07,
+      "step": 2784
+    },
+    {
+      "completion_length": 219.78125762939453,
+      "epoch": 1.4869333333333334,
+      "grad_norm": 0.20918211340904236,
+      "learning_rate": 3e-06,
+      "loss": 0.0606,
+      "policy/loss": 8.460997236170442e-08,
+      "reward": 1.4270833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.7083333432674408,
+      "step": 2785
+    },
+    {
+      "epoch": 1.4874666666666667,
+      "grad_norm": 0.2812556326389313,
+      "learning_rate": 3e-06,
+      "loss": 0.0114,
+      "policy/loss": 6.052978029913447e-08,
+      "step": 2786
+    },
+    {
+      "epoch": 1.488,
+      "grad_norm": 0.23442597687244415,
+      "learning_rate": 3e-06,
+      "loss": 0.0597,
+      "policy/loss": 8.468680334772216e-08,
+      "step": 2787
+    },
+    {
+      "epoch": 1.4885333333333333,
+      "grad_norm": 0.4127173125743866,
+      "learning_rate": 3e-06,
+      "loss": 0.0095,
+      "policy/loss": 6.048135503533558e-08,
+      "step": 2788
+    },
+    {
+      "completion_length": 215.27083587646484,
+      "epoch": 1.4890666666666665,
+      "grad_norm": 0.3401331305503845,
+      "learning_rate": 3e-06,
+      "loss": 0.1369,
+      "policy/loss": -0.029935689768965545,
+      "reward": 1.3177083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.6041666567325592,
+      "step": 2789
+    },
+    {
+      "epoch": 1.4896,
+      "grad_norm": 0.3706362545490265,
+      "learning_rate": 3e-06,
+      "loss": -0.1398,
+      "policy/loss": 0.07665649741939262,
+      "step": 2790
+    },
+    {
+      "epoch": 1.4901333333333333,
+      "grad_norm": 0.36226898431777954,
+      "learning_rate": 3e-06,
+      "loss": 0.1366,
+      "policy/loss": -0.025663994381378785,
+      "step": 2791
+    },
+    {
+      "epoch": 1.4906666666666666,
+      "grad_norm": 0.36963608860969543,
+      "learning_rate": 3e-06,
+      "loss": -0.1411,
+      "policy/loss": 0.07949571990274151,
+      "step": 2792
+    },
+    {
+      "completion_length": 219.6041717529297,
+      "epoch": 1.4912,
+      "grad_norm": 0.2818935215473175,
+      "learning_rate": 3e-06,
+      "loss": -0.0433,
+      "policy/loss": 1.0433865327286185e-07,
+      "reward": 1.4166667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7291666865348816,
+      "step": 2793
+    },
+    {
+      "epoch": 1.4917333333333334,
+      "grad_norm": 0.3103330433368683,
+      "learning_rate": 3e-06,
+      "loss": -0.0321,
+      "policy/loss": 1.1780336706124217e-07,
+      "step": 2794
+    },
+    {
+      "epoch": 1.4922666666666666,
+      "grad_norm": 0.28694021701812744,
+      "learning_rate": 3e-06,
+      "loss": -0.0446,
+      "policy/loss": 1.0430959562768294e-07,
+      "step": 2795
+    },
+    {
+      "epoch": 1.4928,
+      "grad_norm": 0.3117823004722595,
+      "learning_rate": 3e-06,
+      "loss": -0.0323,
+      "policy/loss": 1.1788043607907639e-07,
+      "step": 2796
+    },
+    {
+      "completion_length": 224.3229217529297,
+      "epoch": 1.4933333333333334,
+      "grad_norm": 0.48523348569869995,
+      "learning_rate": 3e-06,
+      "loss": -0.0643,
+      "policy/loss": 4.203267067737215e-08,
+      "reward": 1.3125,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.625,
+      "step": 2797
+    },
+    {
+      "epoch": 1.4938666666666667,
+      "grad_norm": 0.3607572019100189,
+      "learning_rate": 3e-06,
+      "loss": -0.1051,
+      "policy/loss": 4.420147448058742e-08,
+      "step": 2798
+    },
+    {
+      "epoch": 1.4944,
+      "grad_norm": 0.46872949600219727,
+      "learning_rate": 3e-06,
+      "loss": -0.0653,
+      "policy/loss": 4.166771816471737e-08,
+      "step": 2799
+    },
+    {
+      "epoch": 1.4949333333333334,
+      "grad_norm": 0.3573829233646393,
+      "learning_rate": 3e-06,
+      "loss": -0.1076,
+      "policy/loss": 4.422835075956755e-08,
+      "step": 2800
+    },
+    {
+      "completion_length": 217.65625,
+      "epoch": 1.4954666666666667,
+      "grad_norm": 0.3614371716976166,
+      "learning_rate": 3e-06,
+      "loss": -0.0347,
+      "policy/loss": -0.01873255195096135,
+      "reward": 1.4427083730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.7291666567325592,
+      "step": 2801
+    },
+    {
+      "epoch": 1.496,
+      "grad_norm": 0.2618682086467743,
+      "learning_rate": 3e-06,
+      "loss": -0.0173,
+      "policy/loss": -0.1759178787469864,
+      "step": 2802
+    },
+    {
+      "epoch": 1.4965333333333333,
+      "grad_norm": 0.40555548667907715,
+      "learning_rate": 3e-06,
+      "loss": -0.0359,
+      "policy/loss": -0.028584751300513744,
+      "step": 2803
+    },
+    {
+      "epoch": 1.4970666666666665,
+      "grad_norm": 0.26051950454711914,
+      "learning_rate": 3e-06,
+      "loss": -0.0177,
+      "policy/loss": -0.17605478316545486,
+      "step": 2804
+    },
+    {
+      "completion_length": 226.0416717529297,
+      "epoch": 1.4976,
+      "grad_norm": 0.21353089809417725,
+      "learning_rate": 3e-06,
+      "loss": 0.1441,
+      "policy/loss": 0.3998926728963852,
+      "reward": 1.4270833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.7708333283662796,
+      "step": 2805
+    },
+    {
+      "epoch": 1.4981333333333333,
+      "grad_norm": 0.24679729342460632,
+      "learning_rate": 3e-06,
+      "loss": -0.0558,
+      "policy/loss": 0.1423298791050911,
+      "step": 2806
+    },
+    {
+      "epoch": 1.4986666666666666,
+      "grad_norm": 0.20547086000442505,
+      "learning_rate": 3e-06,
+      "loss": 0.144,
+      "policy/loss": 0.39916642010211945,
+      "step": 2807
+    },
+    {
+      "epoch": 1.4992,
+      "grad_norm": 0.25851479172706604,
+      "learning_rate": 3e-06,
+      "loss": -0.0575,
+      "policy/loss": 0.13827326893806458,
+      "step": 2808
+    },
+    {
+      "completion_length": 206.0416717529297,
+      "epoch": 1.4997333333333334,
+      "grad_norm": 0.33717525005340576,
+      "learning_rate": 3e-06,
+      "loss": 0.0162,
+      "policy/loss": 0.3697668449514566,
+      "reward": 1.4843750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.7708333134651184,
+      "step": 2809
+    },
+    {
+      "epoch": 1.5002666666666666,
+      "grad_norm": 0.34887978434562683,
+      "learning_rate": 3e-06,
+      "loss": 0.0815,
+      "policy/loss": 0.09304073512282685,
+      "step": 2810
+    },
+    {
+      "epoch": 1.5008,
+      "grad_norm": 0.36947137117385864,
+      "learning_rate": 3e-06,
+      "loss": 0.0149,
+      "policy/loss": 0.36522693795870254,
+      "step": 2811
+    },
+    {
+      "epoch": 1.5013333333333332,
+      "grad_norm": 0.41842368245124817,
+      "learning_rate": 3e-06,
+      "loss": 0.0791,
+      "policy/loss": 0.0934050908174271,
+      "step": 2812
+    },
+    {
+      "completion_length": 222.4791717529297,
+      "epoch": 1.5018666666666667,
+      "grad_norm": 0.27782443165779114,
+      "learning_rate": 3e-06,
+      "loss": -0.0351,
+      "policy/loss": 7.251272071684411e-08,
+      "reward": 1.5885417461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.875,
+      "step": 2813
+    },
+    {
+      "epoch": 1.5024,
+      "grad_norm": 0.3324369192123413,
+      "learning_rate": 3e-06,
+      "loss": -0.0219,
+      "policy/loss": 6.421792697608453e-08,
+      "step": 2814
+    },
+    {
+      "epoch": 1.5029333333333335,
+      "grad_norm": 0.2902623116970062,
+      "learning_rate": 3e-06,
+      "loss": -0.0359,
+      "policy/loss": 7.253903389070615e-08,
+      "step": 2815
+    },
+    {
+      "epoch": 1.5034666666666667,
+      "grad_norm": 0.3063806891441345,
+      "learning_rate": 3e-06,
+      "loss": -0.0231,
+      "policy/loss": 6.351606973709067e-08,
+      "step": 2816
+    },
+    {
+      "completion_length": 203.14583587646484,
+      "epoch": 1.504,
+      "grad_norm": 0.33492395281791687,
+      "learning_rate": 3e-06,
+      "loss": -0.0638,
+      "policy/loss": 1.6184614537451125e-07,
+      "reward": 1.8072916865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 1.0625,
+      "step": 2817
+    },
+    {
+      "epoch": 1.5045333333333333,
+      "grad_norm": 0.3784602880477905,
+      "learning_rate": 3e-06,
+      "loss": -0.1388,
+      "policy/loss": 1.7831662546541338e-07,
+      "step": 2818
+    },
+    {
+      "epoch": 1.5050666666666666,
+      "grad_norm": 0.31789904832839966,
+      "learning_rate": 3e-06,
+      "loss": -0.0652,
+      "policy/loss": 1.6197240881865582e-07,
+      "step": 2819
+    },
+    {
+      "epoch": 1.5056,
+      "grad_norm": 0.625442385673523,
+      "learning_rate": 3e-06,
+      "loss": -0.1414,
+      "policy/loss": 1.7840258692558564e-07,
+      "step": 2820
+    },
+    {
+      "completion_length": 224.53125762939453,
+      "epoch": 1.5061333333333333,
+      "grad_norm": 0.2692788243293762,
+      "learning_rate": 3e-06,
+      "loss": 0.03,
+      "policy/loss": 0.1131131149449871,
+      "reward": 1.4687500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.8124999701976776,
+      "step": 2821
+    },
+    {
+      "epoch": 1.5066666666666668,
+      "grad_norm": 0.2840917408466339,
+      "learning_rate": 3e-06,
+      "loss": -0.0756,
+      "policy/loss": -0.9011113839280718,
+      "step": 2822
+    },
+    {
+      "epoch": 1.5072,
+      "grad_norm": 0.2725621461868286,
+      "learning_rate": 3e-06,
+      "loss": 0.0274,
+      "policy/loss": 0.10767916816647372,
+      "step": 2823
+    },
+    {
+      "epoch": 1.5077333333333334,
+      "grad_norm": 0.265013188123703,
+      "learning_rate": 3e-06,
+      "loss": -0.0773,
+      "policy/loss": -0.9092316823640232,
+      "step": 2824
+    },
+    {
+      "completion_length": 215.2604217529297,
+      "epoch": 1.5082666666666666,
+      "grad_norm": 0.31159907579421997,
+      "learning_rate": 3e-06,
+      "loss": -0.2012,
+      "policy/loss": -0.23284327376783942,
+      "reward": 1.5937500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8749999701976776,
+      "step": 2825
+    },
+    {
+      "epoch": 1.5088,
+      "grad_norm": 0.2842368185520172,
+      "learning_rate": 3e-06,
+      "loss": -0.1663,
+      "policy/loss": -0.13861395161043788,
+      "step": 2826
+    },
+    {
+      "epoch": 1.5093333333333332,
+      "grad_norm": 0.33418095111846924,
+      "learning_rate": 3e-06,
+      "loss": -0.203,
+      "policy/loss": -0.2336140363297936,
+      "step": 2827
+    },
+    {
+      "epoch": 1.5098666666666667,
+      "grad_norm": 0.29134443402290344,
+      "learning_rate": 3e-06,
+      "loss": -0.1674,
+      "policy/loss": -0.13992640121364275,
+      "step": 2828
+    },
+    {
+      "completion_length": 215.8541717529297,
+      "epoch": 1.5104,
+      "grad_norm": 0.2836761176586151,
+      "learning_rate": 3e-06,
+      "loss": 0.033,
+      "policy/loss": -0.2688797729202861,
+      "reward": 1.9322917461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 1.1875,
+      "step": 2829
+    },
+    {
+      "epoch": 1.5109333333333335,
+      "grad_norm": 0.2989494204521179,
+      "learning_rate": 3e-06,
+      "loss": -0.0161,
+      "policy/loss": -0.183213542489014,
+      "step": 2830
+    },
+    {
+      "epoch": 1.5114666666666667,
+      "grad_norm": 0.3056129217147827,
+      "learning_rate": 3e-06,
+      "loss": 0.0322,
+      "policy/loss": -0.27195701175697096,
+      "step": 2831
+    },
+    {
+      "epoch": 1.512,
+      "grad_norm": 0.2713351845741272,
+      "learning_rate": 3e-06,
+      "loss": -0.0167,
+      "policy/loss": -0.18352898521967376,
+      "step": 2832
+    },
+    {
+      "completion_length": 225.2604217529297,
+      "epoch": 1.5125333333333333,
+      "grad_norm": 0.2569260597229004,
+      "learning_rate": 3e-06,
+      "loss": -0.0173,
+      "policy/loss": 5.7921049290143856e-08,
+      "reward": 1.0885416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.4375,
+      "step": 2833
+    },
+    {
+      "epoch": 1.5130666666666666,
+      "grad_norm": 0.2635868489742279,
+      "learning_rate": 3e-06,
+      "loss": -0.009,
+      "policy/loss": 6.577732136747727e-08,
+      "step": 2834
+    },
+    {
+      "epoch": 1.5135999999999998,
+      "grad_norm": 0.3247238099575043,
+      "learning_rate": 3e-06,
+      "loss": -0.0165,
+      "policy/loss": 5.7970964917331e-08,
+      "step": 2835
+    },
+    {
+      "epoch": 1.5141333333333333,
+      "grad_norm": 0.20498159527778625,
+      "learning_rate": 3e-06,
+      "loss": -0.0097,
+      "policy/loss": 6.576372157951482e-08,
+      "step": 2836
+    },
+    {
+      "completion_length": 203.8541717529297,
+      "epoch": 1.5146666666666668,
+      "grad_norm": 0.26740938425064087,
+      "learning_rate": 3e-06,
+      "loss": -0.2238,
+      "policy/loss": -0.15536813809198335,
+      "reward": 1.484375,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.7708333432674408,
+      "step": 2837
+    },
+    {
+      "epoch": 1.5152,
+      "grad_norm": 0.2581579089164734,
+      "learning_rate": 3e-06,
+      "loss": 0.014,
+      "policy/loss": 0.43759065200410774,
+      "step": 2838
+    },
+    {
+      "epoch": 1.5157333333333334,
+      "grad_norm": 0.27757003903388977,
+      "learning_rate": 3e-06,
+      "loss": -0.2243,
+      "policy/loss": -0.15769948448707538,
+      "step": 2839
+    },
+    {
+      "epoch": 1.5162666666666667,
+      "grad_norm": 0.2742457389831543,
+      "learning_rate": 3e-06,
+      "loss": 0.0133,
+      "policy/loss": 0.43746274040282174,
+      "step": 2840
+    },
+    {
+      "completion_length": 219.78125,
+      "epoch": 1.5168,
+      "grad_norm": 0.27873286604881287,
+      "learning_rate": 3e-06,
+      "loss": -0.1039,
+      "policy/loss": 1.3402636866999273e-07,
+      "reward": 1.4270833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.7083333432674408,
+      "step": 2841
+    },
+    {
+      "epoch": 1.5173333333333332,
+      "grad_norm": 0.2557285726070404,
+      "learning_rate": 3e-06,
+      "loss": -0.0143,
+      "policy/loss": 1.2058807286052797e-07,
+      "step": 2842
+    },
+    {
+      "epoch": 1.5178666666666667,
+      "grad_norm": 0.2857459485530853,
+      "learning_rate": 3e-06,
+      "loss": -0.1044,
+      "policy/loss": 1.3415599298127745e-07,
+      "step": 2843
+    },
+    {
+      "epoch": 1.5184,
+      "grad_norm": 0.25281253457069397,
+      "learning_rate": 3e-06,
+      "loss": -0.0161,
+      "policy/loss": 1.2067303956087017e-07,
+      "step": 2844
+    },
+    {
+      "completion_length": 211.53125762939453,
+      "epoch": 1.5189333333333335,
+      "grad_norm": 0.3317967653274536,
+      "learning_rate": 3e-06,
+      "loss": -0.0012,
+      "policy/loss": 1.1503588837058487e-07,
+      "reward": 1.5312500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8124999701976776,
+      "step": 2845
+    },
+    {
+      "epoch": 1.5194666666666667,
+      "grad_norm": 0.327046662569046,
+      "learning_rate": 3e-06,
+      "loss": -0.0923,
+      "policy/loss": 1.4447812901607904e-07,
+      "step": 2846
+    },
+    {
+      "epoch": 1.52,
+      "grad_norm": 0.3517250120639801,
+      "learning_rate": 3e-06,
+      "loss": -0.0008,
+      "policy/loss": 1.152315043384533e-07,
+      "step": 2847
+    },
+    {
+      "epoch": 1.5205333333333333,
+      "grad_norm": 0.33787041902542114,
+      "learning_rate": 3e-06,
+      "loss": -0.0941,
+      "policy/loss": 1.4464341546727155e-07,
+      "step": 2848
+    },
+    {
+      "completion_length": 203.23958587646484,
+      "epoch": 1.5210666666666666,
+      "grad_norm": 0.26329463720321655,
+      "learning_rate": 3e-06,
+      "loss": 0.1135,
+      "policy/loss": 7.469386709146875e-08,
+      "reward": 1.1770833432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.5208333358168602,
+      "step": 2849
+    },
+    {
+      "epoch": 1.5215999999999998,
+      "grad_norm": 0.2840081453323364,
+      "learning_rate": 3e-06,
+      "loss": -0.0822,
+      "policy/loss": 6.857487822031771e-08,
+      "step": 2850
+    },
+    {
+      "epoch": 1.5221333333333333,
+      "grad_norm": 0.2433900237083435,
+      "learning_rate": 3e-06,
+      "loss": 0.1141,
+      "policy/loss": 7.482968911176613e-08,
+      "step": 2851
+    },
+    {
+      "epoch": 1.5226666666666666,
+      "grad_norm": 0.2764270603656769,
+      "learning_rate": 3e-06,
+      "loss": -0.0839,
+      "policy/loss": 6.874241531562575e-08,
+      "step": 2852
+    },
+    {
+      "completion_length": 205.3854217529297,
+      "epoch": 1.5232,
+      "grad_norm": 0.26217129826545715,
+      "learning_rate": 3e-06,
+      "loss": 0.0677,
+      "policy/loss": 0.4182140864868771,
+      "reward": 1.5104166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7395833432674408,
+      "rewards/correctness_reward_func_math": 0.7708333134651184,
+      "step": 2853
+    },
+    {
+      "epoch": 1.5237333333333334,
+      "grad_norm": 0.20833951234817505,
+      "learning_rate": 3e-06,
+      "loss": 0.0659,
+      "policy/loss": -0.17850506544664313,
+      "step": 2854
+    },
+    {
+      "epoch": 1.5242666666666667,
+      "grad_norm": 0.2646869122982025,
+      "learning_rate": 3e-06,
+      "loss": 0.0655,
+      "policy/loss": 0.4181178546643807,
+      "step": 2855
+    },
+    {
+      "epoch": 1.5248,
+      "grad_norm": 0.239494189620018,
+      "learning_rate": 3e-06,
+      "loss": 0.0669,
+      "policy/loss": -0.17905499281572546,
+      "step": 2856
+    },
+    {
+      "completion_length": 207.27083587646484,
+      "epoch": 1.5253333333333332,
+      "grad_norm": 0.300955206155777,
+      "learning_rate": 3e-06,
+      "loss": -0.0281,
+      "policy/loss": 6.938903140962793e-08,
+      "reward": 1.3281250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.6458333134651184,
+      "step": 2857
+    },
+    {
+      "epoch": 1.5258666666666667,
+      "grad_norm": 0.42354777455329895,
+      "learning_rate": 3e-06,
+      "loss": -0.0625,
+      "policy/loss": 6.601199942224412e-08,
+      "step": 2858
+    },
+    {
+      "epoch": 1.5264,
+      "grad_norm": 0.31283295154571533,
+      "learning_rate": 3e-06,
+      "loss": -0.0292,
+      "policy/loss": 6.951416509082264e-08,
+      "step": 2859
+    },
+    {
+      "epoch": 1.5269333333333335,
+      "grad_norm": 0.3797967731952667,
+      "learning_rate": 3e-06,
+      "loss": -0.0642,
+      "policy/loss": 6.60291625820264e-08,
+      "step": 2860
+    },
+    {
+      "completion_length": 195.0729217529297,
+      "epoch": 1.5274666666666668,
+      "grad_norm": 0.2600419223308563,
+      "learning_rate": 3e-06,
+      "loss": 0.0193,
+      "policy/loss": 1.582588993187528e-07,
+      "reward": 1.4010416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.6875,
+      "step": 2861
+    },
+    {
+      "epoch": 1.528,
+      "grad_norm": 0.2512087821960449,
+      "learning_rate": 3e-06,
+      "loss": -0.0962,
+      "policy/loss": 1.4308054119283042e-07,
+      "step": 2862
+    },
+    {
+      "epoch": 1.5285333333333333,
+      "grad_norm": 0.2745581567287445,
+      "learning_rate": 3e-06,
+      "loss": 0.0194,
+      "policy/loss": 1.5825183652395935e-07,
+      "step": 2863
+    },
+    {
+      "epoch": 1.5290666666666666,
+      "grad_norm": 0.2417854219675064,
+      "learning_rate": 3e-06,
+      "loss": -0.0973,
+      "policy/loss": 1.4287965655057633e-07,
+      "step": 2864
+    },
+    {
+      "completion_length": 193.6666717529297,
+      "epoch": 1.5295999999999998,
+      "grad_norm": 0.46446993947029114,
+      "learning_rate": 3e-06,
+      "loss": -0.0798,
+      "policy/loss": 7.589001516095095e-08,
+      "reward": 1.6822917461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.9375,
+      "step": 2865
+    },
+    {
+      "epoch": 1.5301333333333333,
+      "grad_norm": 0.3018963932991028,
+      "learning_rate": 3e-06,
+      "loss": -0.1059,
+      "policy/loss": 8.38452951512636e-08,
+      "step": 2866
+    },
+    {
+      "epoch": 1.5306666666666666,
+      "grad_norm": 0.38835352659225464,
+      "learning_rate": 3e-06,
+      "loss": -0.0819,
+      "policy/loss": 7.552476155581189e-08,
+      "step": 2867
+    },
+    {
+      "epoch": 1.5312000000000001,
+      "grad_norm": 0.2984137535095215,
+      "learning_rate": 3e-06,
+      "loss": -0.1061,
+      "policy/loss": 8.391989503309105e-08,
+      "step": 2868
+    },
+    {
+      "completion_length": 210.70833587646484,
+      "epoch": 1.5317333333333334,
+      "grad_norm": 0.29248756170272827,
+      "learning_rate": 3e-06,
+      "loss": 0.0016,
+      "policy/loss": 9.300562808789437e-08,
+      "reward": 1.2083333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.4583333283662796,
+      "step": 2869
+    },
+    {
+      "epoch": 1.5322666666666667,
+      "grad_norm": 0.26712700724601746,
+      "learning_rate": 3e-06,
+      "loss": -0.0499,
+      "policy/loss": 1.0144126605382553e-07,
+      "step": 2870
+    },
+    {
+      "epoch": 1.5328,
+      "grad_norm": 0.31417563557624817,
+      "learning_rate": 3e-06,
+      "loss": -0.0007,
+      "policy/loss": 9.288105928817458e-08,
+      "step": 2871
+    },
+    {
+      "epoch": 1.5333333333333332,
+      "grad_norm": 0.32462596893310547,
+      "learning_rate": 3e-06,
+      "loss": -0.0523,
+      "policy/loss": 1.0164909625132168e-07,
+      "step": 2872
+    },
+    {
+      "completion_length": 203.6041717529297,
+      "epoch": 1.5338666666666667,
+      "grad_norm": 0.3549703061580658,
+      "learning_rate": 3e-06,
+      "loss": 0.1112,
+      "policy/loss": 0.4400269311921612,
+      "reward": 1.1770833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.4583333283662796,
+      "step": 2873
+    },
+    {
+      "epoch": 1.5344,
+      "grad_norm": 0.26790764927864075,
+      "learning_rate": 3e-06,
+      "loss": 0.0719,
+      "policy/loss": -0.34809298620006146,
+      "step": 2874
+    },
+    {
+      "epoch": 1.5349333333333335,
+      "grad_norm": 0.3614256978034973,
+      "learning_rate": 3e-06,
+      "loss": 0.1095,
+      "policy/loss": 0.4396752339485843,
+      "step": 2875
+    },
+    {
+      "epoch": 1.5354666666666668,
+      "grad_norm": 0.26690608263015747,
+      "learning_rate": 3e-06,
+      "loss": 0.0702,
+      "policy/loss": -0.34965814464969824,
+      "step": 2876
+    },
+    {
+      "completion_length": 199.28125762939453,
+      "epoch": 1.536,
+      "grad_norm": 0.3479495346546173,
+      "learning_rate": 3e-06,
+      "loss": -0.0286,
+      "policy/loss": 1.611839763882017e-07,
+      "reward": 1.5885417461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.9375000298023224,
+      "step": 2877
+    },
+    {
+      "epoch": 1.5365333333333333,
+      "grad_norm": 0.43801528215408325,
+      "learning_rate": 3e-06,
+      "loss": -0.0088,
+      "policy/loss": 1.8622517217181667e-07,
+      "step": 2878
+    },
+    {
+      "epoch": 1.5370666666666666,
+      "grad_norm": 0.34990352392196655,
+      "learning_rate": 3e-06,
+      "loss": -0.0309,
+      "policy/loss": 1.614510374281508e-07,
+      "step": 2879
+    },
+    {
+      "epoch": 1.5375999999999999,
+      "grad_norm": 0.4055558741092682,
+      "learning_rate": 3e-06,
+      "loss": -0.0112,
+      "policy/loss": 1.863547254288278e-07,
+      "step": 2880
+    },
+    {
+      "completion_length": 208.53125762939453,
+      "epoch": 1.5381333333333334,
+      "grad_norm": 0.3083244562149048,
+      "learning_rate": 3e-06,
+      "loss": -0.0067,
+      "policy/loss": 0.3480639241886969,
+      "reward": 1.7135417461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.9999999701976776,
+      "step": 2881
+    },
+    {
+      "epoch": 1.5386666666666666,
+      "grad_norm": 0.2968754768371582,
+      "learning_rate": 3e-06,
+      "loss": 0.0903,
+      "policy/loss": 0.09481916317973571,
+      "step": 2882
+    },
+    {
+      "epoch": 1.5392000000000001,
+      "grad_norm": 0.2903290390968323,
+      "learning_rate": 3e-06,
+      "loss": -0.0082,
+      "policy/loss": 0.34847552405009097,
+      "step": 2883
+    },
+    {
+      "epoch": 1.5397333333333334,
+      "grad_norm": 0.31745442748069763,
+      "learning_rate": 3e-06,
+      "loss": 0.089,
+      "policy/loss": 0.09426250046765006,
+      "step": 2884
+    },
+    {
+      "completion_length": 202.5625,
+      "epoch": 1.5402666666666667,
+      "grad_norm": 0.2884142994880676,
+      "learning_rate": 3e-06,
+      "loss": -0.0686,
+      "policy/loss": 1.6609379827059456e-07,
+      "reward": 1.5729166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8541666716337204,
+      "step": 2885
+    },
+    {
+      "epoch": 1.5408,
+      "grad_norm": 0.329280823469162,
+      "learning_rate": 3e-06,
+      "loss": 0.1105,
+      "policy/loss": 2.1178923503839542e-07,
+      "step": 2886
+    },
+    {
+      "epoch": 1.5413333333333332,
+      "grad_norm": 0.3104100525379181,
+      "learning_rate": 3e-06,
+      "loss": -0.07,
+      "policy/loss": 1.6591222617989843e-07,
+      "step": 2887
+    },
+    {
+      "epoch": 1.5418666666666667,
+      "grad_norm": 0.3447939157485962,
+      "learning_rate": 3e-06,
+      "loss": 0.1083,
+      "policy/loss": 2.1155436513708992e-07,
+      "step": 2888
+    },
+    {
+      "completion_length": 201.48958587646484,
+      "epoch": 1.5424,
+      "grad_norm": 0.30582931637763977,
+      "learning_rate": 3e-06,
+      "loss": 0.0148,
+      "policy/loss": 0.10974618028249239,
+      "reward": 1.5520833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.8333333432674408,
+      "step": 2889
+    },
+    {
+      "epoch": 1.5429333333333335,
+      "grad_norm": 0.34416744112968445,
+      "learning_rate": 3e-06,
+      "loss": -0.018,
+      "policy/loss": 0.319551789411328,
+      "step": 2890
+    },
+    {
+      "epoch": 1.5434666666666668,
+      "grad_norm": 0.3109111785888672,
+      "learning_rate": 3e-06,
+      "loss": 0.0141,
+      "policy/loss": 0.1089414432563629,
+      "step": 2891
+    },
+    {
+      "epoch": 1.544,
+      "grad_norm": 0.37127411365509033,
+      "learning_rate": 3e-06,
+      "loss": -0.0196,
+      "policy/loss": 0.32354139658374237,
+      "step": 2892
+    },
+    {
+      "completion_length": 205.8854217529297,
+      "epoch": 1.5445333333333333,
+      "grad_norm": 0.34336113929748535,
+      "learning_rate": 3e-06,
+      "loss": 0.0379,
+      "policy/loss": -0.1683824633857789,
+      "reward": 1.6718750596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7135416567325592,
+      "rewards/correctness_reward_func_math": 0.9583333134651184,
+      "step": 2893
+    },
+    {
+      "epoch": 1.5450666666666666,
+      "grad_norm": 0.2772943079471588,
+      "learning_rate": 3e-06,
+      "loss": 0.0908,
+      "policy/loss": 0.11717098595476827,
+      "step": 2894
+    },
+    {
+      "epoch": 1.5455999999999999,
+      "grad_norm": 0.33929452300071716,
+      "learning_rate": 3e-06,
+      "loss": 0.036,
+      "policy/loss": -0.1689668720491646,
+      "step": 2895
+    },
+    {
+      "epoch": 1.5461333333333334,
+      "grad_norm": 0.296848863363266,
+      "learning_rate": 3e-06,
+      "loss": 0.089,
+      "policy/loss": 0.11583838260430745,
+      "step": 2896
+    },
+    {
+      "completion_length": 206.14583587646484,
+      "epoch": 1.5466666666666666,
+      "grad_norm": 0.36362361907958984,
+      "learning_rate": 3e-06,
+      "loss": 0.0016,
+      "policy/loss": -0.08257396563609021,
+      "reward": 1.5520833730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6770833432674408,
+      "rewards/correctness_reward_func_math": 0.8749999701976776,
+      "step": 2897
+    },
+    {
+      "epoch": 1.5472000000000001,
+      "grad_norm": 0.23908133804798126,
+      "learning_rate": 3e-06,
+      "loss": -0.0522,
+      "policy/loss": 0.022343699031662112,
+      "step": 2898
+    },
+    {
+      "epoch": 1.5477333333333334,
+      "grad_norm": 0.3585667610168457,
+      "learning_rate": 3e-06,
+      "loss": 0.0004,
+      "policy/loss": -0.08261217964190593,
+      "step": 2899
+    },
+    {
+      "epoch": 1.5482666666666667,
+      "grad_norm": 0.33972257375717163,
+      "learning_rate": 3e-06,
+      "loss": -0.0524,
+      "policy/loss": 0.020452914942815426,
+      "step": 2900
+    },
+    {
+      "completion_length": 206.65625,
+      "epoch": 1.5488,
+      "grad_norm": 0.25341904163360596,
+      "learning_rate": 3e-06,
+      "loss": -0.0686,
+      "policy/loss": -0.5739055052399635,
+      "reward": 1.4635416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6927083432674408,
+      "rewards/correctness_reward_func_math": 0.7708333134651184,
+      "step": 2901
+    },
+    {
+      "epoch": 1.5493333333333332,
+      "grad_norm": 0.24652810394763947,
+      "learning_rate": 3e-06,
+      "loss": -0.1358,
+      "policy/loss": -0.6600601971149445,
+      "step": 2902
+    },
+    {
+      "epoch": 1.5498666666666665,
+      "grad_norm": 0.40614956617355347,
+      "learning_rate": 3e-06,
+      "loss": -0.0696,
+      "policy/loss": -0.5737781450152397,
+      "step": 2903
+    },
+    {
+      "epoch": 1.5504,
+      "grad_norm": 0.25503021478652954,
+      "learning_rate": 3e-06,
+      "loss": -0.137,
+      "policy/loss": -0.6628457754850388,
+      "step": 2904
+    },
+    {
+      "completion_length": 210.5416717529297,
+      "epoch": 1.5509333333333335,
+      "grad_norm": 0.2655560374259949,
+      "learning_rate": 3e-06,
+      "loss": -0.0304,
+      "policy/loss": 0.4160440025877321,
+      "reward": 1.5208333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.8333333134651184,
+      "step": 2905
+    },
+    {
+      "epoch": 1.5514666666666668,
+      "grad_norm": 0.2736262381076813,
+      "learning_rate": 3e-06,
+      "loss": 0.1642,
+      "policy/loss": 0.5062594195901653,
+      "step": 2906
+    },
+    {
+      "epoch": 1.552,
+      "grad_norm": 0.271602600812912,
+      "learning_rate": 3e-06,
+      "loss": -0.0318,
+      "policy/loss": 0.40836656660306403,
+      "step": 2907
+    },
+    {
+      "epoch": 1.5525333333333333,
+      "grad_norm": 0.2739487588405609,
+      "learning_rate": 3e-06,
+      "loss": 0.1633,
+      "policy/loss": 0.5021672626032156,
+      "step": 2908
+    },
+    {
+      "completion_length": 210.03125,
+      "epoch": 1.5530666666666666,
+      "grad_norm": 0.2808612287044525,
+      "learning_rate": 3e-06,
+      "loss": -0.0107,
+      "policy/loss": -0.03564523905515671,
+      "reward": 1.3489583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6822916567325592,
+      "rewards/correctness_reward_func_math": 0.6666666865348816,
+      "step": 2909
+    },
+    {
+      "epoch": 1.5535999999999999,
+      "grad_norm": 0.33325961232185364,
+      "learning_rate": 3e-06,
+      "loss": -0.0738,
+      "policy/loss": -0.2731311097741127,
+      "step": 2910
+    },
+    {
+      "epoch": 1.5541333333333334,
+      "grad_norm": 0.3205932378768921,
+      "learning_rate": 3e-06,
+      "loss": -0.0119,
+      "policy/loss": -0.04501807689666748,
+      "step": 2911
+    },
+    {
+      "epoch": 1.5546666666666666,
+      "grad_norm": 0.3305119574069977,
+      "learning_rate": 3e-06,
+      "loss": -0.0743,
+      "policy/loss": -0.2750483565032482,
+      "step": 2912
+    },
+    {
+      "completion_length": 202.37500762939453,
+      "epoch": 1.5552000000000001,
+      "grad_norm": 0.35538700222969055,
+      "learning_rate": 3e-06,
+      "loss": 0.0169,
+      "policy/loss": 1.489221510553307e-07,
+      "reward": 1.4583333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.8333333432674408,
+      "step": 2913
+    },
+    {
+      "epoch": 1.5557333333333334,
+      "grad_norm": 0.2703152298927307,
+      "learning_rate": 3e-06,
+      "loss": 0.0133,
+      "policy/loss": 1.5495555061306732e-07,
+      "step": 2914
+    },
+    {
+      "epoch": 1.5562666666666667,
+      "grad_norm": 0.3591960668563843,
+      "learning_rate": 3e-06,
+      "loss": 0.015,
+      "policy/loss": 1.4889646848814664e-07,
+      "step": 2915
+    },
+    {
+      "epoch": 1.5568,
+      "grad_norm": 0.26404696702957153,
+      "learning_rate": 3e-06,
+      "loss": 0.0125,
+      "policy/loss": 1.5502710226655836e-07,
+      "step": 2916
+    },
+    {
+      "completion_length": 206.8229217529297,
+      "epoch": 1.5573333333333332,
+      "grad_norm": 0.24266262352466583,
+      "learning_rate": 3e-06,
+      "loss": 0.1109,
+      "policy/loss": 5.997536689505978e-08,
+      "reward": 1.2187500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.59375,
+      "rewards/correctness_reward_func_math": 0.6249999850988388,
+      "step": 2917
+    },
+    {
+      "epoch": 1.5578666666666665,
+      "grad_norm": 0.23471087217330933,
+      "learning_rate": 3e-06,
+      "loss": 0.137,
+      "policy/loss": 6.833825949570382e-08,
+      "step": 2918
+    },
+    {
+      "epoch": 1.5584,
+      "grad_norm": 0.2315884232521057,
+      "learning_rate": 3e-06,
+      "loss": 0.111,
+      "policy/loss": 5.999391028410628e-08,
+      "step": 2919
+    },
+    {
+      "epoch": 1.5589333333333333,
+      "grad_norm": 0.26323530077934265,
+      "learning_rate": 3e-06,
+      "loss": 0.1365,
+      "policy/loss": 6.836474852889296e-08,
+      "step": 2920
+    },
+    {
+      "completion_length": 206.17708587646484,
+      "epoch": 1.5594666666666668,
+      "grad_norm": 0.2768736779689789,
+      "learning_rate": 3e-06,
+      "loss": -0.1412,
+      "policy/loss": 0.0697718196035293,
+      "reward": 1.6458333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.8958333432674408,
+      "step": 2921
+    },
+    {
+      "epoch": 1.56,
+      "grad_norm": 0.30463463068008423,
+      "learning_rate": 3e-06,
+      "loss": -0.0422,
+      "policy/loss": 0.01416605661907333,
+      "step": 2922
+    },
+    {
+      "epoch": 1.5605333333333333,
+      "grad_norm": 0.28701484203338623,
+      "learning_rate": 3e-06,
+      "loss": -0.1414,
+      "policy/loss": 0.07083021185071026,
+      "step": 2923
+    },
+    {
+      "epoch": 1.5610666666666666,
+      "grad_norm": 0.3916301131248474,
+      "learning_rate": 3e-06,
+      "loss": -0.0441,
+      "policy/loss": 0.015194206948045519,
+      "step": 2924
+    },
+    {
+      "completion_length": 208.62500762939453,
+      "epoch": 1.5615999999999999,
+      "grad_norm": 0.2696510851383209,
+      "learning_rate": 3e-06,
+      "loss": 0.099,
+      "policy/loss": 0.15551646105854644,
+      "reward": 1.3958333730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7083333134651184,
+      "step": 2925
+    },
+    {
+      "epoch": 1.5621333333333334,
+      "grad_norm": 0.2562619745731354,
+      "learning_rate": 3e-06,
+      "loss": -0.0369,
+      "policy/loss": -0.31349488929181035,
+      "step": 2926
+    },
+    {
+      "epoch": 1.5626666666666666,
+      "grad_norm": 0.25894540548324585,
+      "learning_rate": 3e-06,
+      "loss": 0.0974,
+      "policy/loss": 0.15592023774788544,
+      "step": 2927
+    },
+    {
+      "epoch": 1.5632000000000001,
+      "grad_norm": 0.23732425272464752,
+      "learning_rate": 3e-06,
+      "loss": -0.0381,
+      "policy/loss": -0.32115515897964997,
+      "step": 2928
+    },
+    {
+      "completion_length": 197.23959350585938,
+      "epoch": 1.5637333333333334,
+      "grad_norm": 0.4392601251602173,
+      "learning_rate": 3e-06,
+      "loss": -0.0873,
+      "policy/loss": 1.2461068621405502e-07,
+      "reward": 1.6822917461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.9375,
+      "step": 2929
+    },
+    {
+      "epoch": 1.5642666666666667,
+      "grad_norm": 0.34427955746650696,
+      "learning_rate": 3e-06,
+      "loss": 0.0301,
+      "policy/loss": 1.1551895795491873e-07,
+      "step": 2930
+    },
+    {
+      "epoch": 1.5648,
+      "grad_norm": 0.4113837480545044,
+      "learning_rate": 3e-06,
+      "loss": -0.0877,
+      "policy/loss": 1.2457437037483032e-07,
+      "step": 2931
+    },
+    {
+      "epoch": 1.5653333333333332,
+      "grad_norm": 0.3808252215385437,
+      "learning_rate": 3e-06,
+      "loss": 0.0274,
+      "policy/loss": 1.1532707233641304e-07,
+      "step": 2932
+    },
+    {
+      "completion_length": 211.8854217529297,
+      "epoch": 1.5658666666666665,
+      "grad_norm": 0.22970212996006012,
+      "learning_rate": 3e-06,
+      "loss": 0.1083,
+      "policy/loss": 0.6832134276628494,
+      "reward": 1.0364583432674408,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6197916567325592,
+      "rewards/correctness_reward_func_math": 0.4166666716337204,
+      "step": 2933
+    },
+    {
+      "epoch": 1.5664,
+      "grad_norm": 0.2910361886024475,
+      "learning_rate": 3e-06,
+      "loss": 0.0641,
+      "policy/loss": 0.33377429842948914,
+      "step": 2934
+    },
+    {
+      "epoch": 1.5669333333333333,
+      "grad_norm": 0.24231868982315063,
+      "learning_rate": 3e-06,
+      "loss": 0.1075,
+      "policy/loss": 0.6772793382406235,
+      "step": 2935
+    },
+    {
+      "epoch": 1.5674666666666668,
+      "grad_norm": 0.2504938542842865,
+      "learning_rate": 3e-06,
+      "loss": 0.0636,
+      "policy/loss": 0.32974426448345184,
+      "step": 2936
+    },
+    {
+      "completion_length": 188.25000762939453,
+      "epoch": 1.568,
+      "grad_norm": 0.2526865601539612,
+      "learning_rate": 3e-06,
+      "loss": 0.0331,
+      "policy/loss": 1.113531062912898e-07,
+      "reward": 1.5937500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6770833432674408,
+      "rewards/correctness_reward_func_math": 0.9166666567325592,
+      "step": 2937
+    },
+    {
+      "epoch": 1.5685333333333333,
+      "grad_norm": 0.5294008255004883,
+      "learning_rate": 3e-06,
+      "loss": -0.0241,
+      "policy/loss": 1.0478533951641111e-07,
+      "step": 2938
+    },
+    {
+      "epoch": 1.5690666666666666,
+      "grad_norm": 0.24808436632156372,
+      "learning_rate": 3e-06,
+      "loss": 0.0327,
+      "policy/loss": 1.1118863696424341e-07,
+      "step": 2939
+    },
+    {
+      "epoch": 1.5695999999999999,
+      "grad_norm": 0.3349294066429138,
+      "learning_rate": 3e-06,
+      "loss": -0.0257,
+      "policy/loss": 1.0477809908593372e-07,
+      "step": 2940
+    },
+    {
+      "completion_length": 215.28125762939453,
+      "epoch": 1.5701333333333334,
+      "grad_norm": 0.40966272354125977,
+      "learning_rate": 3e-06,
+      "loss": 0.0364,
+      "policy/loss": 1.145117778378335e-07,
+      "reward": 1.4166667461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.666666641831398,
+      "step": 2941
+    },
+    {
+      "epoch": 1.5706666666666667,
+      "grad_norm": 0.34929323196411133,
+      "learning_rate": 3e-06,
+      "loss": 0.14,
+      "policy/loss": 1.12265119867061e-07,
+      "step": 2942
+    },
+    {
+      "epoch": 1.5712000000000002,
+      "grad_norm": 0.3997287452220917,
+      "learning_rate": 3e-06,
+      "loss": 0.034,
+      "policy/loss": 1.1444915060110361e-07,
+      "step": 2943
+    },
+    {
+      "epoch": 1.5717333333333334,
+      "grad_norm": 0.334077924489975,
+      "learning_rate": 3e-06,
+      "loss": 0.1372,
+      "policy/loss": 1.121558881322926e-07,
+      "step": 2944
+    },
+    {
+      "completion_length": 201.80208587646484,
+      "epoch": 1.5722666666666667,
+      "grad_norm": 0.26867929100990295,
+      "learning_rate": 3e-06,
+      "loss": 0.0071,
+      "policy/loss": 1.4911905665826453e-07,
+      "reward": 1.1875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.75,
+      "rewards/correctness_reward_func_math": 0.4375,
+      "step": 2945
+    },
+    {
+      "epoch": 1.5728,
+      "grad_norm": 0.3646833896636963,
+      "learning_rate": 3e-06,
+      "loss": -0.0514,
+      "policy/loss": 1.5223730542857083e-07,
+      "step": 2946
+    },
+    {
+      "epoch": 1.5733333333333333,
+      "grad_norm": 0.2537921667098999,
+      "learning_rate": 3e-06,
+      "loss": 0.0056,
+      "policy/loss": 1.491718819579546e-07,
+      "step": 2947
+    },
+    {
+      "epoch": 1.5738666666666665,
+      "grad_norm": 0.363135427236557,
+      "learning_rate": 3e-06,
+      "loss": -0.053,
+      "policy/loss": 1.522708785728355e-07,
+      "step": 2948
+    },
+    {
+      "completion_length": 211.9166717529297,
+      "epoch": 1.5744,
+      "grad_norm": 0.29848039150238037,
+      "learning_rate": 3e-06,
+      "loss": -0.0314,
+      "policy/loss": 6.067492108741135e-08,
+      "reward": 1.5104166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.8541666567325592,
+      "step": 2949
+    },
+    {
+      "epoch": 1.5749333333333333,
+      "grad_norm": 0.2812748849391937,
+      "learning_rate": 3e-06,
+      "loss": -0.1366,
+      "policy/loss": 5.919488188510513e-08,
+      "step": 2950
+    },
+    {
+      "epoch": 1.5754666666666668,
+      "grad_norm": 0.2861405611038208,
+      "learning_rate": 3e-06,
+      "loss": -0.0326,
+      "policy/loss": 6.067193680792116e-08,
+      "step": 2951
+    },
+    {
+      "epoch": 1.576,
+      "grad_norm": 0.2942732870578766,
+      "learning_rate": 3e-06,
+      "loss": -0.1393,
+      "policy/loss": 5.90669699818136e-08,
+      "step": 2952
+    },
+    {
+      "completion_length": 217.7291717529297,
+      "epoch": 1.5765333333333333,
+      "grad_norm": 0.3310207426548004,
+      "learning_rate": 3e-06,
+      "loss": 0.0021,
+      "policy/loss": -0.14384452604504716,
+      "reward": 1.1875000298023224,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.4999999850988388,
+      "step": 2953
+    },
+    {
+      "epoch": 1.5770666666666666,
+      "grad_norm": 0.4014424979686737,
+      "learning_rate": 3e-06,
+      "loss": 0.1558,
+      "policy/loss": 0.0823123594469628,
+      "step": 2954
+    },
+    {
+      "epoch": 1.5776,
+      "grad_norm": 0.31060293316841125,
+      "learning_rate": 3e-06,
+      "loss": 0.0006,
+      "policy/loss": -0.14519650848014365,
+      "step": 2955
+    },
+    {
+      "epoch": 1.5781333333333334,
+      "grad_norm": 0.40571853518486023,
+      "learning_rate": 3e-06,
+      "loss": 0.1513,
+      "policy/loss": 0.08176255872547955,
+      "step": 2956
+    },
+    {
+      "completion_length": 213.90625762939453,
+      "epoch": 1.5786666666666667,
+      "grad_norm": 0.4913206994533539,
+      "learning_rate": 3e-06,
+      "loss": -0.0941,
+      "policy/loss": -0.23413533353721583,
+      "reward": 1.5260417461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.8749999701976776,
+      "step": 2957
+    },
+    {
+      "epoch": 1.5792000000000002,
+      "grad_norm": 0.2436683028936386,
+      "learning_rate": 3e-06,
+      "loss": 0.0639,
+      "policy/loss": 0.22420075135742668,
+      "step": 2958
+    },
+    {
+      "epoch": 1.5797333333333334,
+      "grad_norm": 0.3683513402938843,
+      "learning_rate": 3e-06,
+      "loss": -0.0947,
+      "policy/loss": -0.23565215254095673,
+      "step": 2959
+    },
+    {
+      "epoch": 1.5802666666666667,
+      "grad_norm": 0.2733921706676483,
+      "learning_rate": 3e-06,
+      "loss": 0.0627,
+      "policy/loss": 0.2225199599805201,
+      "step": 2960
+    },
+    {
+      "completion_length": 214.30209350585938,
+      "epoch": 1.5808,
+      "grad_norm": 0.3247927725315094,
+      "learning_rate": 3e-06,
+      "loss": 0.1818,
+      "policy/loss": 0.04009712194781656,
+      "reward": 1.4687500596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.75,
+      "step": 2961
+    },
+    {
+      "epoch": 1.5813333333333333,
+      "grad_norm": 0.3367268145084381,
+      "learning_rate": 3e-06,
+      "loss": 0.1196,
+      "policy/loss": -0.3575497892552342,
+      "step": 2962
+    },
+    {
+      "epoch": 1.5818666666666665,
+      "grad_norm": 0.352765828371048,
+      "learning_rate": 3e-06,
+      "loss": 0.1803,
+      "policy/loss": 0.033956118833330606,
+      "step": 2963
+    },
+    {
+      "epoch": 1.5824,
+      "grad_norm": 0.3752710223197937,
+      "learning_rate": 3e-06,
+      "loss": 0.1164,
+      "policy/loss": -0.3568241324887502,
+      "step": 2964
+    },
+    {
+      "completion_length": 214.68750762939453,
+      "epoch": 1.5829333333333333,
+      "grad_norm": 0.2691033184528351,
+      "learning_rate": 3e-06,
+      "loss": -0.0728,
+      "policy/loss": 9.837984649152531e-08,
+      "reward": 1.4479167461395264,
+      "rewards/boxed_and_answer_tags_format_reward": 0.71875,
+      "rewards/correctness_reward_func_math": 0.7291666865348816,
+      "step": 2965
+    },
+    {
+      "epoch": 1.5834666666666668,
+      "grad_norm": 0.41842904686927795,
+      "learning_rate": 3e-06,
+      "loss": 0.1894,
+      "policy/loss": 1.1549479594918921e-07,
+      "step": 2966
+    },
+    {
+      "epoch": 1.584,
+      "grad_norm": 0.2707744538784027,
+      "learning_rate": 3e-06,
+      "loss": -0.0733,
+      "policy/loss": 9.818899471270015e-08,
+      "step": 2967
+    },
+    {
+      "epoch": 1.5845333333333333,
+      "grad_norm": 0.3985583484172821,
+      "learning_rate": 3e-06,
+      "loss": 0.1892,
+      "policy/loss": 1.1539546562744363e-07,
+      "step": 2968
+    },
+    {
+      "completion_length": 219.20833587646484,
+      "epoch": 1.5850666666666666,
+      "grad_norm": 0.36127710342407227,
+      "learning_rate": 3e-06,
+      "loss": 0.157,
+      "policy/loss": 0.5728115439414978,
+      "reward": 1.6614583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.9166666567325592,
+      "step": 2969
+    },
+    {
+      "epoch": 1.5856,
+      "grad_norm": 0.3735817074775696,
+      "learning_rate": 3e-06,
+      "loss": 0.0379,
+      "policy/loss": -0.18775498121976852,
+      "step": 2970
+    },
+    {
+      "epoch": 1.5861333333333332,
+      "grad_norm": 0.37353891134262085,
+      "learning_rate": 3e-06,
+      "loss": 0.1539,
+      "policy/loss": 0.5699783563613892,
+      "step": 2971
+    },
+    {
+      "epoch": 1.5866666666666667,
+      "grad_norm": 0.35461562871932983,
+      "learning_rate": 3e-06,
+      "loss": 0.0359,
+      "policy/loss": -0.18582724407315254,
+      "step": 2972
+    },
+    {
+      "completion_length": 217.14583587646484,
+      "epoch": 1.5872000000000002,
+      "grad_norm": 0.24860087037086487,
+      "learning_rate": 3e-06,
+      "loss": 0.0239,
+      "policy/loss": -0.003918049595085904,
+      "reward": 1.4166666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.7916666716337204,
+      "step": 2973
+    },
+    {
+      "epoch": 1.5877333333333334,
+      "grad_norm": 0.29557958245277405,
+      "learning_rate": 3e-06,
+      "loss": -0.2445,
+      "policy/loss": -0.34554941952228546,
+      "step": 2974
+    },
+    {
+      "epoch": 1.5882666666666667,
+      "grad_norm": 0.2989253103733063,
+      "learning_rate": 3e-06,
+      "loss": 0.0232,
+      "policy/loss": -0.005505824112333357,
+      "step": 2975
+    },
+    {
+      "epoch": 1.5888,
+      "grad_norm": 0.307435005903244,
+      "learning_rate": 3e-06,
+      "loss": -0.2466,
+      "policy/loss": -0.34986579418182373,
+      "step": 2976
+    },
+    {
+      "completion_length": 216.0729217529297,
+      "epoch": 1.5893333333333333,
+      "grad_norm": 0.2192407101392746,
+      "learning_rate": 3e-06,
+      "loss": -0.1921,
+      "policy/loss": -0.6053874492645264,
+      "reward": 1.4791666865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6875,
+      "rewards/correctness_reward_func_math": 0.7916666567325592,
+      "step": 2977
+    },
+    {
+      "epoch": 1.5898666666666665,
+      "grad_norm": 0.22923307120800018,
+      "learning_rate": 3e-06,
+      "loss": -0.0101,
+      "policy/loss": 0.2550448589026928,
+      "step": 2978
+    },
+    {
+      "epoch": 1.5904,
+      "grad_norm": 0.2218644917011261,
+      "learning_rate": 3e-06,
+      "loss": -0.1926,
+      "policy/loss": -0.6051739752292633,
+      "step": 2979
+    },
+    {
+      "epoch": 1.5909333333333333,
+      "grad_norm": 0.2515302002429962,
+      "learning_rate": 3e-06,
+      "loss": -0.0119,
+      "policy/loss": 0.25211823731660843,
+      "step": 2980
+    },
+    {
+      "completion_length": 223.80209350585938,
+      "epoch": 1.5914666666666668,
+      "grad_norm": 0.33332744240760803,
+      "learning_rate": 3e-06,
+      "loss": -0.0564,
+      "policy/loss": 1.1142003586428473e-07,
+      "reward": 1.2135416865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.6510416567325592,
+      "rewards/correctness_reward_func_math": 0.5625000149011612,
+      "step": 2981
+    },
+    {
+      "epoch": 1.592,
+      "grad_norm": 0.3775920569896698,
+      "learning_rate": 3e-06,
+      "loss": -0.0386,
+      "policy/loss": 9.969212300120489e-08,
+      "step": 2982
+    },
+    {
+      "epoch": 1.5925333333333334,
+      "grad_norm": 0.3100719451904297,
+      "learning_rate": 3e-06,
+      "loss": -0.0567,
+      "policy/loss": 1.1144461709022835e-07,
+      "step": 2983
+    },
+    {
+      "epoch": 1.5930666666666666,
+      "grad_norm": 0.3516411781311035,
+      "learning_rate": 3e-06,
+      "loss": -0.0397,
+      "policy/loss": 9.968800540605116e-08,
+      "step": 2984
+    },
+    {
+      "completion_length": 223.03125762939453,
+      "epoch": 1.5936,
+      "grad_norm": 0.2166157066822052,
+      "learning_rate": 3e-06,
+      "loss": -0.0921,
+      "policy/loss": 0.357308606140613,
+      "reward": 1.1875,
+      "rewards/boxed_and_answer_tags_format_reward": 0.625,
+      "rewards/correctness_reward_func_math": 0.5625,
+      "step": 2985
+    },
+    {
+      "epoch": 1.5941333333333332,
+      "grad_norm": 0.2567504346370697,
+      "learning_rate": 3e-06,
+      "loss": -0.1194,
+      "policy/loss": -0.19082788143109042,
+      "step": 2986
+    },
+    {
+      "epoch": 1.5946666666666667,
+      "grad_norm": 0.21473826467990875,
+      "learning_rate": 3e-06,
+      "loss": -0.093,
+      "policy/loss": 0.3556820847772286,
+      "step": 2987
+    },
+    {
+      "epoch": 1.5952,
+      "grad_norm": 0.25862911343574524,
+      "learning_rate": 3e-06,
+      "loss": -0.1201,
+      "policy/loss": -0.19199875494904717,
+      "step": 2988
+    },
+    {
+      "completion_length": 202.5104217529297,
+      "epoch": 1.5957333333333334,
+      "grad_norm": 0.2762909233570099,
+      "learning_rate": 3e-06,
+      "loss": 0.0533,
+      "policy/loss": 1.334790944440556e-07,
+      "reward": 1.0364583730697632,
+      "rewards/boxed_and_answer_tags_format_reward": 0.640625,
+      "rewards/correctness_reward_func_math": 0.3958333283662796,
+      "step": 2989
+    },
+    {
+      "epoch": 1.5962666666666667,
+      "grad_norm": 0.2527987062931061,
+      "learning_rate": 3e-06,
+      "loss": 0.0257,
+      "policy/loss": 1.411659908967522e-07,
+      "step": 2990
+    },
+    {
+      "epoch": 1.5968,
+      "grad_norm": 0.28133222460746765,
+      "learning_rate": 3e-06,
+      "loss": 0.0519,
+      "policy/loss": 1.3311700186591224e-07,
+      "step": 2991
+    },
+    {
+      "epoch": 1.5973333333333333,
+      "grad_norm": 0.26959750056266785,
+      "learning_rate": 3e-06,
+      "loss": 0.0255,
+      "policy/loss": 1.409251524364663e-07,
+      "step": 2992
+    },
+    {
+      "completion_length": 232.39584350585938,
+      "epoch": 1.5978666666666665,
+      "grad_norm": 0.5492620468139648,
+      "learning_rate": 3e-06,
+      "loss": -0.2084,
+      "policy/loss": -0.09395767026245494,
+      "reward": 1.3281250596046448,
+      "rewards/boxed_and_answer_tags_format_reward": 0.7447916567325592,
+      "rewards/correctness_reward_func_math": 0.5833333432674408,
+      "step": 2993
+    },
+    {
+      "epoch": 1.5984,
+      "grad_norm": 0.34232011437416077,
+      "learning_rate": 3e-06,
+      "loss": 0.1422,
+      "policy/loss": 0.15909029614833514,
+      "step": 2994
+    },
+    {
+      "epoch": 1.5989333333333333,
+      "grad_norm": 0.38131529092788696,
+      "learning_rate": 3e-06,
+      "loss": -0.2085,
+      "policy/loss": -0.0975045861485242,
+      "step": 2995
+    },
+    {
+      "epoch": 1.5994666666666668,
+      "grad_norm": 0.30277687311172485,
+      "learning_rate": 3e-06,
+      "loss": 0.1409,
+      "policy/loss": 0.1570356793918748,
+      "step": 2996
+    },
+    {
+      "completion_length": 215.05208587646484,
+      "epoch": 1.6,
+      "grad_norm": 0.27154073119163513,
+      "learning_rate": 3e-06,
+      "loss": -0.0904,
+      "policy/loss": 1.0067646627476279e-07,
+      "reward": 1.1354166865348816,
+      "rewards/boxed_and_answer_tags_format_reward": 0.65625,
+      "rewards/correctness_reward_func_math": 0.4791666716337204,
+      "step": 2997
+    },
+    {
+      "epoch": 1.6005333333333334,
+      "grad_norm": 0.1991499960422516,
+      "learning_rate": 3e-06,
+      "loss": 0.0485,
+      "policy/loss": 8.710136611966846e-08,
+      "step": 2998
+    },
+    {
+      "epoch": 1.6010666666666666,
+      "grad_norm": 0.29584378004074097,
+      "learning_rate": 3e-06,
+      "loss": -0.0918,
+      "policy/loss": 1.0065064515174527e-07,
+      "step": 2999
+    },
+    {
+      "epoch": 1.6016,
+      "grad_norm": 0.1958972066640854,
+      "learning_rate": 3e-06,
+      "loss": 0.0475,
+      "policy/loss": 8.701394094146053e-08,
+      "step": 3000
+    }
+  ],
+  "logging_steps": 1,
+  "max_steps": 5625,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 3,
+  "save_steps": 100,
+  "stateful_callbacks": {
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 0.0,
+  "train_batch_size": 6,
+  "trial_name": null,
+  "trial_params": null
+}