diff --git "a/checkpoint-214/trainer_state.json" "b/checkpoint-214/trainer_state.json"
--- "a/checkpoint-214/trainer_state.json"
+++ "b/checkpoint-214/trainer_state.json"
@@ -1,7 +1,7 @@
 {
   "best_metric": null,
   "best_model_checkpoint": null,
-  "epoch": 1.2081863091037404,
+  "epoch": 1.202247191011236,
   "eval_steps": 27,
   "global_step": 214,
   "is_hyper_param_search": false,
@@ -9,1534 +9,1534 @@
   "is_world_process_zero": true,
   "log_history": [
     {
-      "epoch": 0.016937191249117856,
-      "grad_norm": 34.22002029418945,
-      "learning_rate": 6.818181818181818e-07,
-      "loss": 7.2372,
+      "epoch": 0.016853932584269662,
+      "grad_norm": 2.9885776042938232,
+      "learning_rate": 1.4814814814814815e-06,
+      "loss": 0.6012,
       "step": 3
     },
     {
-      "epoch": 0.03387438249823571,
-      "grad_norm": 21.76839828491211,
-      "learning_rate": 1.3636363636363636e-06,
-      "loss": 6.855,
+      "epoch": 0.033707865168539325,
+      "grad_norm": 3.184929132461548,
+      "learning_rate": 2.962962962962963e-06,
+      "loss": 0.7573,
       "step": 6
     },
     {
-      "epoch": 0.05081157374735357,
-      "grad_norm": 21.260774612426758,
-      "learning_rate": 2.0454545454545457e-06,
-      "loss": 7.4707,
+      "epoch": 0.05056179775280899,
+      "grad_norm": 3.256159782409668,
+      "learning_rate": 4.444444444444444e-06,
+      "loss": 0.9212,
       "step": 9
     },
     {
-      "epoch": 0.06774876499647142,
-      "grad_norm": 16.885921478271484,
-      "learning_rate": 2.7272727272727272e-06,
-      "loss": 7.0187,
+      "epoch": 0.06741573033707865,
+      "grad_norm": 2.833339214324951,
+      "learning_rate": 5.925925925925926e-06,
+      "loss": 0.6117,
       "step": 12
     },
     {
-      "epoch": 0.08468595624558928,
-      "grad_norm": 19.509899139404297,
-      "learning_rate": 3.409090909090909e-06,
-      "loss": 6.6756,
+      "epoch": 0.08426966292134831,
+      "grad_norm": 3.08292818069458,
+      "learning_rate": 7.4074074074074075e-06,
+      "loss": 0.8545,
       "step": 15
     },
     {
-      "epoch": 0.10162314749470713,
-      "grad_norm": 7.9427289962768555,
-      "learning_rate": 4.0909090909090915e-06,
-      "loss": 6.0155,
+      "epoch": 0.10112359550561797,
+      "grad_norm": 2.317431688308716,
+      "learning_rate": 8.888888888888888e-06,
+      "loss": 0.6515,
       "step": 18
     },
     {
-      "epoch": 0.11856033874382499,
-      "grad_norm": 7.325345039367676,
-      "learning_rate": 4.772727272727273e-06,
-      "loss": 6.1644,
+      "epoch": 0.11797752808988764,
+      "grad_norm": 2.9611644744873047,
+      "learning_rate": 1.037037037037037e-05,
+      "loss": 0.7159,
       "step": 21
     },
     {
-      "epoch": 0.13549752999294284,
-      "grad_norm": 7.544689655303955,
-      "learning_rate": 5.4545454545454545e-06,
-      "loss": 6.2158,
+      "epoch": 0.1348314606741573,
+      "grad_norm": 2.698537826538086,
+      "learning_rate": 1.1851851851851852e-05,
+      "loss": 0.7019,
       "step": 24
     },
     {
-      "epoch": 0.1524347212420607,
-      "grad_norm": 5.141758918762207,
-      "learning_rate": 6.136363636363637e-06,
-      "loss": 6.1369,
+      "epoch": 0.15168539325842698,
+      "grad_norm": 2.222154378890991,
+      "learning_rate": 1.3333333333333333e-05,
+      "loss": 0.4411,
       "step": 27
     },
     {
-      "epoch": 0.1524347212420607,
+      "epoch": 0.15168539325842698,
       "eval_NLI-v2_cosine_accuracy": 1.0,
-      "eval_NLI-v2_dot_accuracy": 0.109375,
+      "eval_NLI-v2_dot_accuracy": 0.0,
       "eval_NLI-v2_euclidean_accuracy": 1.0,
       "eval_NLI-v2_manhattan_accuracy": 1.0,
       "eval_NLI-v2_max_accuracy": 1.0,
-      "eval_VitaminC_cosine_accuracy": 0.5546875,
-      "eval_VitaminC_cosine_accuracy_threshold": 0.9544724822044373,
-      "eval_VitaminC_cosine_ap": 0.5356492030729136,
-      "eval_VitaminC_cosine_f1": 0.6542553191489362,
-      "eval_VitaminC_cosine_f1_threshold": 0.7148199081420898,
-      "eval_VitaminC_cosine_precision": 0.48616600790513836,
+      "eval_VitaminC_cosine_accuracy": 0.578125,
+      "eval_VitaminC_cosine_accuracy_threshold": 0.7817381620407104,
+      "eval_VitaminC_cosine_ap": 0.5507972943944112,
+      "eval_VitaminC_cosine_f1": 0.6595174262734584,
+      "eval_VitaminC_cosine_f1_threshold": 0.28573715686798096,
+      "eval_VitaminC_cosine_precision": 0.492,
       "eval_VitaminC_cosine_recall": 1.0,
-      "eval_VitaminC_dot_accuracy": 0.55078125,
-      "eval_VitaminC_dot_accuracy_threshold": 414.4264831542969,
-      "eval_VitaminC_dot_ap": 0.5108219546857565,
-      "eval_VitaminC_dot_f1": 0.6507936507936508,
-      "eval_VitaminC_dot_f1_threshold": 271.6522521972656,
-      "eval_VitaminC_dot_precision": 0.4823529411764706,
+      "eval_VitaminC_dot_accuracy": 0.5703125,
+      "eval_VitaminC_dot_accuracy_threshold": 316.7283020019531,
+      "eval_VitaminC_dot_ap": 0.5511866185449577,
+      "eval_VitaminC_dot_f1": 0.6577540106951871,
+      "eval_VitaminC_dot_f1_threshold": 106.75863647460938,
+      "eval_VitaminC_dot_precision": 0.4900398406374502,
       "eval_VitaminC_dot_recall": 1.0,
-      "eval_VitaminC_euclidean_accuracy": 0.55078125,
-      "eval_VitaminC_euclidean_accuracy_threshold": 6.519885063171387,
-      "eval_VitaminC_euclidean_ap": 0.5226419655984281,
-      "eval_VitaminC_euclidean_f1": 0.6505376344086021,
-      "eval_VitaminC_euclidean_f1_threshold": 15.194067001342773,
-      "eval_VitaminC_euclidean_precision": 0.4859437751004016,
-      "eval_VitaminC_euclidean_recall": 0.983739837398374,
-      "eval_VitaminC_manhattan_accuracy": 0.546875,
-      "eval_VitaminC_manhattan_accuracy_threshold": 149.20114135742188,
-      "eval_VitaminC_manhattan_ap": 0.5237451656134715,
-      "eval_VitaminC_manhattan_f1": 0.6542553191489362,
-      "eval_VitaminC_manhattan_f1_threshold": 259.007080078125,
-      "eval_VitaminC_manhattan_precision": 0.48616600790513836,
+      "eval_VitaminC_euclidean_accuracy": 0.578125,
+      "eval_VitaminC_euclidean_accuracy_threshold": 13.298419952392578,
+      "eval_VitaminC_euclidean_ap": 0.5476323986807207,
+      "eval_VitaminC_euclidean_f1": 0.6577540106951871,
+      "eval_VitaminC_euclidean_f1_threshold": 23.83933448791504,
+      "eval_VitaminC_euclidean_precision": 0.4900398406374502,
+      "eval_VitaminC_euclidean_recall": 1.0,
+      "eval_VitaminC_manhattan_accuracy": 0.578125,
+      "eval_VitaminC_manhattan_accuracy_threshold": 279.69085693359375,
+      "eval_VitaminC_manhattan_ap": 0.5412538781107805,
+      "eval_VitaminC_manhattan_f1": 0.6577540106951871,
+      "eval_VitaminC_manhattan_f1_threshold": 499.8836364746094,
+      "eval_VitaminC_manhattan_precision": 0.4900398406374502,
       "eval_VitaminC_manhattan_recall": 1.0,
-      "eval_VitaminC_max_accuracy": 0.5546875,
-      "eval_VitaminC_max_accuracy_threshold": 414.4264831542969,
-      "eval_VitaminC_max_ap": 0.5356492030729136,
-      "eval_VitaminC_max_f1": 0.6542553191489362,
-      "eval_VitaminC_max_f1_threshold": 271.6522521972656,
-      "eval_VitaminC_max_precision": 0.48616600790513836,
+      "eval_VitaminC_max_accuracy": 0.578125,
+      "eval_VitaminC_max_accuracy_threshold": 316.7283020019531,
+      "eval_VitaminC_max_ap": 0.5511866185449577,
+      "eval_VitaminC_max_f1": 0.6595174262734584,
+      "eval_VitaminC_max_f1_threshold": 499.8836364746094,
+      "eval_VitaminC_max_precision": 0.492,
       "eval_VitaminC_max_recall": 1.0,
-      "eval_sequential_score": 0.5356492030729136,
-      "eval_sts-test_pearson_cosine": 0.056062031998983373,
-      "eval_sts-test_pearson_dot": 0.2979259445723872,
-      "eval_sts-test_pearson_euclidean": 0.0498319208592713,
-      "eval_sts-test_pearson_manhattan": 0.07381429239121526,
-      "eval_sts-test_pearson_max": 0.2979259445723872,
-      "eval_sts-test_spearman_cosine": 0.1066788491614481,
-      "eval_sts-test_spearman_dot": 0.315952670306405,
-      "eval_sts-test_spearman_euclidean": 0.07303394554435191,
-      "eval_sts-test_spearman_manhattan": 0.09039525717692232,
-      "eval_sts-test_spearman_max": 0.315952670306405,
-      "eval_vitaminc-pairs_loss": 2.698580741882324,
-      "eval_vitaminc-pairs_runtime": 1.4747,
-      "eval_vitaminc-pairs_samples_per_second": 73.236,
-      "eval_vitaminc-pairs_steps_per_second": 1.356,
+      "eval_sequential_score": 0.5511866185449577,
+      "eval_sts-test_pearson_cosine": 0.8488243436029344,
+      "eval_sts-test_pearson_dot": 0.8480167969551653,
+      "eval_sts-test_pearson_euclidean": 0.8800283985117625,
+      "eval_sts-test_pearson_manhattan": 0.880588311422627,
+      "eval_sts-test_pearson_max": 0.880588311422627,
+      "eval_sts-test_spearman_cosine": 0.8905659331642088,
+      "eval_sts-test_spearman_dot": 0.8692084657204004,
+      "eval_sts-test_spearman_euclidean": 0.8809566840232712,
+      "eval_sts-test_spearman_manhattan": 0.883434007028195,
+      "eval_sts-test_spearman_max": 0.8905659331642088,
+      "eval_vitaminc-pairs_loss": 2.465860366821289,
+      "eval_vitaminc-pairs_runtime": 1.4615,
+      "eval_vitaminc-pairs_samples_per_second": 73.899,
+      "eval_vitaminc-pairs_steps_per_second": 1.368,
       "step": 27
     },
     {
-      "epoch": 0.1524347212420607,
-      "eval_negation-triplets_loss": 5.142906665802002,
-      "eval_negation-triplets_runtime": 0.2993,
-      "eval_negation-triplets_samples_per_second": 213.865,
-      "eval_negation-triplets_steps_per_second": 3.342,
+      "epoch": 0.15168539325842698,
+      "eval_negation-triplets_loss": 1.7310789823532104,
+      "eval_negation-triplets_runtime": 0.3009,
+      "eval_negation-triplets_samples_per_second": 212.692,
+      "eval_negation-triplets_steps_per_second": 3.323,
       "step": 27
     },
     {
-      "epoch": 0.1524347212420607,
-      "eval_scitail-pairs-pos_loss": 1.9216958284378052,
-      "eval_scitail-pairs-pos_runtime": 0.3834,
-      "eval_scitail-pairs-pos_samples_per_second": 140.842,
-      "eval_scitail-pairs-pos_steps_per_second": 2.608,
+      "epoch": 0.15168539325842698,
+      "eval_scitail-pairs-pos_loss": 0.1150394082069397,
+      "eval_scitail-pairs-pos_runtime": 0.3739,
+      "eval_scitail-pairs-pos_samples_per_second": 144.431,
+      "eval_scitail-pairs-pos_steps_per_second": 2.675,
       "step": 27
     },
     {
-      "epoch": 0.1524347212420607,
-      "eval_xsum-pairs_loss": 6.073049545288086,
-      "eval_xsum-pairs_runtime": 3.1587,
-      "eval_xsum-pairs_samples_per_second": 40.523,
-      "eval_xsum-pairs_steps_per_second": 0.633,
+      "epoch": 0.15168539325842698,
+      "eval_xsum-pairs_loss": 0.11168850213289261,
+      "eval_xsum-pairs_runtime": 3.1697,
+      "eval_xsum-pairs_samples_per_second": 40.382,
+      "eval_xsum-pairs_steps_per_second": 0.631,
       "step": 27
     },
     {
-      "epoch": 0.1524347212420607,
-      "eval_sciq_pairs_loss": 0.3449864387512207,
-      "eval_sciq_pairs_runtime": 3.3747,
-      "eval_sciq_pairs_samples_per_second": 37.93,
-      "eval_sciq_pairs_steps_per_second": 0.593,
+      "epoch": 0.15168539325842698,
+      "eval_sciq_pairs_loss": 0.03450964391231537,
+      "eval_sciq_pairs_runtime": 3.3283,
+      "eval_sciq_pairs_samples_per_second": 38.459,
+      "eval_sciq_pairs_steps_per_second": 0.601,
       "step": 27
     },
     {
-      "epoch": 0.1524347212420607,
-      "eval_qasc_pairs_loss": 3.2267842292785645,
-      "eval_qasc_pairs_runtime": 0.6576,
-      "eval_qasc_pairs_samples_per_second": 194.646,
-      "eval_qasc_pairs_steps_per_second": 3.041,
+      "epoch": 0.15168539325842698,
+      "eval_qasc_pairs_loss": 0.11095743626356125,
+      "eval_qasc_pairs_runtime": 0.6261,
+      "eval_qasc_pairs_samples_per_second": 204.45,
+      "eval_qasc_pairs_steps_per_second": 3.195,
       "step": 27
     },
     {
-      "epoch": 0.1524347212420607,
-      "eval_openbookqa_pairs_loss": 4.405983924865723,
-      "eval_openbookqa_pairs_runtime": 0.6107,
-      "eval_openbookqa_pairs_samples_per_second": 209.594,
-      "eval_openbookqa_pairs_steps_per_second": 3.275,
+      "epoch": 0.15168539325842698,
+      "eval_openbookqa_pairs_loss": 0.7092063426971436,
+      "eval_openbookqa_pairs_runtime": 0.5866,
+      "eval_openbookqa_pairs_samples_per_second": 218.19,
+      "eval_openbookqa_pairs_steps_per_second": 3.409,
       "step": 27
     },
     {
-      "epoch": 0.1524347212420607,
-      "eval_msmarco_pairs_loss": 6.937691688537598,
-      "eval_msmarco_pairs_runtime": 1.3091,
-      "eval_msmarco_pairs_samples_per_second": 97.779,
-      "eval_msmarco_pairs_steps_per_second": 1.528,
+      "epoch": 0.15168539325842698,
+      "eval_msmarco_pairs_loss": 0.3955218493938446,
+      "eval_msmarco_pairs_runtime": 1.2942,
+      "eval_msmarco_pairs_samples_per_second": 98.902,
+      "eval_msmarco_pairs_steps_per_second": 1.545,
       "step": 27
     },
     {
-      "epoch": 0.1524347212420607,
-      "eval_nq_pairs_loss": 6.794108867645264,
-      "eval_nq_pairs_runtime": 2.3968,
-      "eval_nq_pairs_samples_per_second": 53.404,
-      "eval_nq_pairs_steps_per_second": 0.834,
+      "epoch": 0.15168539325842698,
+      "eval_nq_pairs_loss": 0.42051073908805847,
+      "eval_nq_pairs_runtime": 2.3875,
+      "eval_nq_pairs_samples_per_second": 53.612,
+      "eval_nq_pairs_steps_per_second": 0.838,
       "step": 27
     },
     {
-      "epoch": 0.1524347212420607,
-      "eval_trivia_pairs_loss": 6.3355631828308105,
-      "eval_trivia_pairs_runtime": 4.4974,
-      "eval_trivia_pairs_samples_per_second": 28.461,
-      "eval_trivia_pairs_steps_per_second": 0.445,
+      "epoch": 0.15168539325842698,
+      "eval_trivia_pairs_loss": 0.93178790807724,
+      "eval_trivia_pairs_runtime": 4.4363,
+      "eval_trivia_pairs_samples_per_second": 28.853,
+      "eval_trivia_pairs_steps_per_second": 0.451,
       "step": 27
     },
     {
-      "epoch": 0.1524347212420607,
-      "eval_gooaq_pairs_loss": 6.405998706817627,
-      "eval_gooaq_pairs_runtime": 0.8745,
-      "eval_gooaq_pairs_samples_per_second": 146.37,
-      "eval_gooaq_pairs_steps_per_second": 2.287,
+      "epoch": 0.15168539325842698,
+      "eval_gooaq_pairs_loss": 0.6505913138389587,
+      "eval_gooaq_pairs_runtime": 0.8826,
+      "eval_gooaq_pairs_samples_per_second": 145.027,
+      "eval_gooaq_pairs_steps_per_second": 2.266,
       "step": 27
     },
     {
-      "epoch": 0.1524347212420607,
-      "eval_paws-pos_loss": 2.2308223247528076,
-      "eval_paws-pos_runtime": 0.6998,
-      "eval_paws-pos_samples_per_second": 182.908,
-      "eval_paws-pos_steps_per_second": 2.858,
+      "epoch": 0.15168539325842698,
+      "eval_paws-pos_loss": 0.024931101128458977,
+      "eval_paws-pos_runtime": 0.6852,
+      "eval_paws-pos_samples_per_second": 186.805,
+      "eval_paws-pos_steps_per_second": 2.919,
       "step": 27
     },
     {
-      "epoch": 0.16937191249117856,
-      "grad_norm": 5.885251522064209,
-      "learning_rate": 6.818181818181818e-06,
-      "loss": 5.7653,
+      "epoch": 0.16853932584269662,
+      "grad_norm": 2.826900005340576,
+      "learning_rate": 1.4814814814814815e-05,
+      "loss": 0.5125,
       "step": 30
     },
     {
-      "epoch": 0.1863091037402964,
-      "grad_norm": 7.357480049133301,
-      "learning_rate": 7.500000000000001e-06,
-      "loss": 6.1259,
+      "epoch": 0.1853932584269663,
+      "grad_norm": 2.9938910007476807,
+      "learning_rate": 1.6296296296296297e-05,
+      "loss": 0.6885,
       "step": 33
     },
     {
-      "epoch": 0.20324629498941427,
-      "grad_norm": 7.321795463562012,
-      "learning_rate": 8.181818181818183e-06,
-      "loss": 5.7539,
+      "epoch": 0.20224719101123595,
+      "grad_norm": 3.3046395778656006,
+      "learning_rate": 1.7777777777777777e-05,
+      "loss": 0.6435,
       "step": 36
     },
     {
-      "epoch": 0.22018348623853212,
-      "grad_norm": 4.239792346954346,
-      "learning_rate": 8.863636363636365e-06,
-      "loss": 6.0131,
+      "epoch": 0.21910112359550563,
+      "grad_norm": 2.4184651374816895,
+      "learning_rate": 1.925925925925926e-05,
+      "loss": 0.753,
       "step": 39
     },
     {
-      "epoch": 0.23712067748764998,
-      "grad_norm": 3.9554407596588135,
-      "learning_rate": 9.545454545454547e-06,
-      "loss": 6.0074,
+      "epoch": 0.23595505617977527,
+      "grad_norm": 2.9905433654785156,
+      "learning_rate": 2.074074074074074e-05,
+      "loss": 0.7427,
       "step": 42
     },
     {
-      "epoch": 0.25405786873676783,
-      "grad_norm": 4.406026840209961,
-      "learning_rate": 1.0227272727272729e-05,
-      "loss": 5.7125,
+      "epoch": 0.25280898876404495,
+      "grad_norm": 2.745820999145508,
+      "learning_rate": 2.2222222222222227e-05,
+      "loss": 0.5083,
       "step": 45
     },
     {
-      "epoch": 0.2709950599858857,
-      "grad_norm": 7.235893249511719,
-      "learning_rate": 1.0909090909090909e-05,
-      "loss": 5.5634,
+      "epoch": 0.2696629213483146,
+      "grad_norm": 2.6370577812194824,
+      "learning_rate": 2.3703703703703703e-05,
+      "loss": 0.7454,
       "step": 48
     },
     {
-      "epoch": 0.28793225123500354,
-      "grad_norm": 5.330288410186768,
-      "learning_rate": 1.1590909090909093e-05,
-      "loss": 5.2924,
+      "epoch": 0.28651685393258425,
+      "grad_norm": 3.044011116027832,
+      "learning_rate": 2.5185185185185187e-05,
+      "loss": 0.8356,
       "step": 51
     },
     {
-      "epoch": 0.3048694424841214,
-      "grad_norm": 7.216403961181641,
-      "learning_rate": 1.2272727272727274e-05,
-      "loss": 5.2286,
+      "epoch": 0.30337078651685395,
+      "grad_norm": 3.718804121017456,
+      "learning_rate": 2.6666666666666667e-05,
+      "loss": 0.8864,
       "step": 54
     },
     {
-      "epoch": 0.3048694424841214,
+      "epoch": 0.30337078651685395,
       "eval_NLI-v2_cosine_accuracy": 1.0,
-      "eval_NLI-v2_dot_accuracy": 0.046875,
+      "eval_NLI-v2_dot_accuracy": 0.0,
       "eval_NLI-v2_euclidean_accuracy": 1.0,
       "eval_NLI-v2_manhattan_accuracy": 1.0,
       "eval_NLI-v2_max_accuracy": 1.0,
-      "eval_VitaminC_cosine_accuracy": 0.54296875,
-      "eval_VitaminC_cosine_accuracy_threshold": 0.9328227043151855,
-      "eval_VitaminC_cosine_ap": 0.5212059026196154,
-      "eval_VitaminC_cosine_f1": 0.6576819407008085,
-      "eval_VitaminC_cosine_f1_threshold": 0.7373804450035095,
-      "eval_VitaminC_cosine_precision": 0.49193548387096775,
-      "eval_VitaminC_cosine_recall": 0.991869918699187,
-      "eval_VitaminC_dot_accuracy": 0.55078125,
-      "eval_VitaminC_dot_accuracy_threshold": 418.2774658203125,
-      "eval_VitaminC_dot_ap": 0.5160594099493883,
-      "eval_VitaminC_dot_f1": 0.6521739130434782,
-      "eval_VitaminC_dot_f1_threshold": 291.5081481933594,
-      "eval_VitaminC_dot_precision": 0.4897959183673469,
-      "eval_VitaminC_dot_recall": 0.975609756097561,
-      "eval_VitaminC_euclidean_accuracy": 0.5390625,
-      "eval_VitaminC_euclidean_accuracy_threshold": 8.120429039001465,
-      "eval_VitaminC_euclidean_ap": 0.5224837623095228,
-      "eval_VitaminC_euclidean_f1": 0.6576819407008085,
-      "eval_VitaminC_euclidean_f1_threshold": 14.879999160766602,
-      "eval_VitaminC_euclidean_precision": 0.49193548387096775,
-      "eval_VitaminC_euclidean_recall": 0.991869918699187,
-      "eval_VitaminC_manhattan_accuracy": 0.53515625,
-      "eval_VitaminC_manhattan_accuracy_threshold": 137.40658569335938,
-      "eval_VitaminC_manhattan_ap": 0.5186382518671783,
-      "eval_VitaminC_manhattan_f1": 0.6576086956521738,
-      "eval_VitaminC_manhattan_f1_threshold": 263.32452392578125,
-      "eval_VitaminC_manhattan_precision": 0.49387755102040815,
-      "eval_VitaminC_manhattan_recall": 0.983739837398374,
-      "eval_VitaminC_max_accuracy": 0.55078125,
-      "eval_VitaminC_max_accuracy_threshold": 418.2774658203125,
-      "eval_VitaminC_max_ap": 0.5224837623095228,
-      "eval_VitaminC_max_f1": 0.6576819407008085,
-      "eval_VitaminC_max_f1_threshold": 291.5081481933594,
-      "eval_VitaminC_max_precision": 0.49387755102040815,
-      "eval_VitaminC_max_recall": 0.991869918699187,
-      "eval_sequential_score": 0.5224837623095228,
-      "eval_sts-test_pearson_cosine": 0.14377091128453176,
-      "eval_sts-test_pearson_dot": 0.24728387094758872,
-      "eval_sts-test_pearson_euclidean": 0.14604155960515372,
-      "eval_sts-test_pearson_manhattan": 0.1446467532231986,
-      "eval_sts-test_pearson_max": 0.24728387094758872,
-      "eval_sts-test_spearman_cosine": 0.1968510434344728,
-      "eval_sts-test_spearman_dot": 0.29467218283745694,
-      "eval_sts-test_spearman_euclidean": 0.17218164683969664,
-      "eval_sts-test_spearman_manhattan": 0.17741843340856742,
-      "eval_sts-test_spearman_max": 0.29467218283745694,
-      "eval_vitaminc-pairs_loss": 2.664700746536255,
-      "eval_vitaminc-pairs_runtime": 1.4487,
-      "eval_vitaminc-pairs_samples_per_second": 74.551,
-      "eval_vitaminc-pairs_steps_per_second": 1.381,
+      "eval_VitaminC_cosine_accuracy": 0.57421875,
+      "eval_VitaminC_cosine_accuracy_threshold": 0.7991844415664673,
+      "eval_VitaminC_cosine_ap": 0.5485498837322925,
+      "eval_VitaminC_cosine_f1": 0.6595174262734584,
+      "eval_VitaminC_cosine_f1_threshold": 0.3160865008831024,
+      "eval_VitaminC_cosine_precision": 0.492,
+      "eval_VitaminC_cosine_recall": 1.0,
+      "eval_VitaminC_dot_accuracy": 0.578125,
+      "eval_VitaminC_dot_accuracy_threshold": 327.0416564941406,
+      "eval_VitaminC_dot_ap": 0.54993134882601,
+      "eval_VitaminC_dot_f1": 0.6595174262734584,
+      "eval_VitaminC_dot_f1_threshold": 117.44181060791016,
+      "eval_VitaminC_dot_precision": 0.492,
+      "eval_VitaminC_dot_recall": 1.0,
+      "eval_VitaminC_euclidean_accuracy": 0.57421875,
+      "eval_VitaminC_euclidean_accuracy_threshold": 13.019258499145508,
+      "eval_VitaminC_euclidean_ap": 0.5435066540334542,
+      "eval_VitaminC_euclidean_f1": 0.6577540106951871,
+      "eval_VitaminC_euclidean_f1_threshold": 23.688644409179688,
+      "eval_VitaminC_euclidean_precision": 0.4900398406374502,
+      "eval_VitaminC_euclidean_recall": 1.0,
+      "eval_VitaminC_manhattan_accuracy": 0.57421875,
+      "eval_VitaminC_manhattan_accuracy_threshold": 283.876220703125,
+      "eval_VitaminC_manhattan_ap": 0.5416615397828658,
+      "eval_VitaminC_manhattan_f1": 0.6559999999999999,
+      "eval_VitaminC_manhattan_f1_threshold": 514.0216064453125,
+      "eval_VitaminC_manhattan_precision": 0.4880952380952381,
+      "eval_VitaminC_manhattan_recall": 1.0,
+      "eval_VitaminC_max_accuracy": 0.578125,
+      "eval_VitaminC_max_accuracy_threshold": 327.0416564941406,
+      "eval_VitaminC_max_ap": 0.54993134882601,
+      "eval_VitaminC_max_f1": 0.6595174262734584,
+      "eval_VitaminC_max_f1_threshold": 514.0216064453125,
+      "eval_VitaminC_max_precision": 0.492,
+      "eval_VitaminC_max_recall": 1.0,
+      "eval_sequential_score": 0.54993134882601,
+      "eval_sts-test_pearson_cosine": 0.8452615878553369,
+      "eval_sts-test_pearson_dot": 0.8404858620687519,
+      "eval_sts-test_pearson_euclidean": 0.8780527810910925,
+      "eval_sts-test_pearson_manhattan": 0.878916157345712,
+      "eval_sts-test_pearson_max": 0.878916157345712,
+      "eval_sts-test_spearman_cosine": 0.8876915367075635,
+      "eval_sts-test_spearman_dot": 0.8608104875327304,
+      "eval_sts-test_spearman_euclidean": 0.8804138856889071,
+      "eval_sts-test_spearman_manhattan": 0.8822803815444743,
+      "eval_sts-test_spearman_max": 0.8876915367075635,
+      "eval_vitaminc-pairs_loss": 2.454524040222168,
+      "eval_vitaminc-pairs_runtime": 1.4583,
+      "eval_vitaminc-pairs_samples_per_second": 74.057,
+      "eval_vitaminc-pairs_steps_per_second": 1.371,
       "step": 54
     },
     {
-      "epoch": 0.3048694424841214,
-      "eval_negation-triplets_loss": 4.6218037605285645,
-      "eval_negation-triplets_runtime": 0.2971,
-      "eval_negation-triplets_samples_per_second": 215.438,
-      "eval_negation-triplets_steps_per_second": 3.366,
+      "epoch": 0.30337078651685395,
+      "eval_negation-triplets_loss": 1.7277792692184448,
+      "eval_negation-triplets_runtime": 0.3027,
+      "eval_negation-triplets_samples_per_second": 211.436,
+      "eval_negation-triplets_steps_per_second": 3.304,
       "step": 54
     },
     {
-      "epoch": 0.3048694424841214,
-      "eval_scitail-pairs-pos_loss": 1.2413936853408813,
-      "eval_scitail-pairs-pos_runtime": 0.372,
-      "eval_scitail-pairs-pos_samples_per_second": 145.175,
-      "eval_scitail-pairs-pos_steps_per_second": 2.688,
+      "epoch": 0.30337078651685395,
+      "eval_scitail-pairs-pos_loss": 0.11168555170297623,
+      "eval_scitail-pairs-pos_runtime": 0.3726,
+      "eval_scitail-pairs-pos_samples_per_second": 144.911,
+      "eval_scitail-pairs-pos_steps_per_second": 2.684,
       "step": 54
     },
     {
-      "epoch": 0.3048694424841214,
-      "eval_xsum-pairs_loss": 5.249766826629639,
-      "eval_xsum-pairs_runtime": 3.1506,
-      "eval_xsum-pairs_samples_per_second": 40.627,
-      "eval_xsum-pairs_steps_per_second": 0.635,
+      "epoch": 0.30337078651685395,
+      "eval_xsum-pairs_loss": 0.10087604075670242,
+      "eval_xsum-pairs_runtime": 3.1701,
+      "eval_xsum-pairs_samples_per_second": 40.377,
+      "eval_xsum-pairs_steps_per_second": 0.631,
       "step": 54
     },
     {
-      "epoch": 0.3048694424841214,
-      "eval_sciq_pairs_loss": 0.2961578667163849,
-      "eval_sciq_pairs_runtime": 3.2909,
-      "eval_sciq_pairs_samples_per_second": 38.895,
-      "eval_sciq_pairs_steps_per_second": 0.608,
+      "epoch": 0.30337078651685395,
+      "eval_sciq_pairs_loss": 0.03466618433594704,
+      "eval_sciq_pairs_runtime": 3.3778,
+      "eval_sciq_pairs_samples_per_second": 37.895,
+      "eval_sciq_pairs_steps_per_second": 0.592,
       "step": 54
     },
     {
-      "epoch": 0.3048694424841214,
-      "eval_qasc_pairs_loss": 2.530872344970703,
-      "eval_qasc_pairs_runtime": 0.6255,
-      "eval_qasc_pairs_samples_per_second": 204.63,
-      "eval_qasc_pairs_steps_per_second": 3.197,
+      "epoch": 0.30337078651685395,
+      "eval_qasc_pairs_loss": 0.10551701486110687,
+      "eval_qasc_pairs_runtime": 0.6271,
+      "eval_qasc_pairs_samples_per_second": 204.125,
+      "eval_qasc_pairs_steps_per_second": 3.189,
       "step": 54
     },
     {
-      "epoch": 0.3048694424841214,
-      "eval_openbookqa_pairs_loss": 3.8855104446411133,
-      "eval_openbookqa_pairs_runtime": 0.5742,
-      "eval_openbookqa_pairs_samples_per_second": 222.914,
-      "eval_openbookqa_pairs_steps_per_second": 3.483,
+      "epoch": 0.30337078651685395,
+      "eval_openbookqa_pairs_loss": 0.7239958643913269,
+      "eval_openbookqa_pairs_runtime": 0.5811,
+      "eval_openbookqa_pairs_samples_per_second": 220.255,
+      "eval_openbookqa_pairs_steps_per_second": 3.441,
       "step": 54
     },
     {
-      "epoch": 0.3048694424841214,
-      "eval_msmarco_pairs_loss": 5.246406555175781,
-      "eval_msmarco_pairs_runtime": 1.2872,
-      "eval_msmarco_pairs_samples_per_second": 99.442,
-      "eval_msmarco_pairs_steps_per_second": 1.554,
+      "epoch": 0.30337078651685395,
+      "eval_msmarco_pairs_loss": 0.3808779716491699,
+      "eval_msmarco_pairs_runtime": 1.2919,
+      "eval_msmarco_pairs_samples_per_second": 99.082,
+      "eval_msmarco_pairs_steps_per_second": 1.548,
       "step": 54
     },
     {
-      "epoch": 0.3048694424841214,
-      "eval_nq_pairs_loss": 5.332630157470703,
-      "eval_nq_pairs_runtime": 2.3739,
-      "eval_nq_pairs_samples_per_second": 53.92,
-      "eval_nq_pairs_steps_per_second": 0.843,
+      "epoch": 0.30337078651685395,
+      "eval_nq_pairs_loss": 0.44170400500297546,
+      "eval_nq_pairs_runtime": 2.3835,
+      "eval_nq_pairs_samples_per_second": 53.703,
+      "eval_nq_pairs_steps_per_second": 0.839,
       "step": 54
     },
     {
-      "epoch": 0.3048694424841214,
-      "eval_trivia_pairs_loss": 5.647429943084717,
-      "eval_trivia_pairs_runtime": 4.4729,
-      "eval_trivia_pairs_samples_per_second": 28.617,
-      "eval_trivia_pairs_steps_per_second": 0.447,
+      "epoch": 0.30337078651685395,
+      "eval_trivia_pairs_loss": 0.9158428907394409,
+      "eval_trivia_pairs_runtime": 4.4326,
+      "eval_trivia_pairs_samples_per_second": 28.877,
+      "eval_trivia_pairs_steps_per_second": 0.451,
       "step": 54
     },
     {
-      "epoch": 0.3048694424841214,
-      "eval_gooaq_pairs_loss": 5.225871562957764,
-      "eval_gooaq_pairs_runtime": 0.8715,
-      "eval_gooaq_pairs_samples_per_second": 146.868,
-      "eval_gooaq_pairs_steps_per_second": 2.295,
+      "epoch": 0.30337078651685395,
+      "eval_gooaq_pairs_loss": 0.6208247542381287,
+      "eval_gooaq_pairs_runtime": 0.8797,
+      "eval_gooaq_pairs_samples_per_second": 145.497,
+      "eval_gooaq_pairs_steps_per_second": 2.273,
       "step": 54
     },
     {
-      "epoch": 0.3048694424841214,
-      "eval_paws-pos_loss": 0.8335962891578674,
-      "eval_paws-pos_runtime": 0.6844,
-      "eval_paws-pos_samples_per_second": 187.036,
-      "eval_paws-pos_steps_per_second": 2.922,
+      "epoch": 0.30337078651685395,
+      "eval_paws-pos_loss": 0.02517784759402275,
+      "eval_paws-pos_runtime": 0.694,
+      "eval_paws-pos_samples_per_second": 184.442,
+      "eval_paws-pos_steps_per_second": 2.882,
       "step": 54
     },
     {
-      "epoch": 0.32180663373323926,
-      "grad_norm": 6.847682952880859,
-      "learning_rate": 1.2954545454545455e-05,
-      "loss": 4.4811,
+      "epoch": 0.3202247191011236,
+      "grad_norm": 2.173736572265625,
+      "learning_rate": 2.814814814814815e-05,
+      "loss": 0.6015,
       "step": 57
     },
     {
-      "epoch": 0.3387438249823571,
-      "grad_norm": 8.383002281188965,
-      "learning_rate": 1.3636363636363637e-05,
-      "loss": 4.4239,
+      "epoch": 0.33707865168539325,
+      "grad_norm": 3.8964712619781494,
+      "learning_rate": 2.962962962962963e-05,
+      "loss": 0.9482,
       "step": 60
     },
     {
-      "epoch": 0.35568101623147497,
-      "grad_norm": 7.014843463897705,
-      "learning_rate": 1.431818181818182e-05,
-      "loss": 4.0273,
+      "epoch": 0.3539325842696629,
+      "grad_norm": 2.659498691558838,
+      "learning_rate": 3.111111111111112e-05,
+      "loss": 0.5404,
       "step": 63
     },
     {
-      "epoch": 0.3726182074805928,
-      "grad_norm": 5.9739885330200195,
-      "learning_rate": 1.5000000000000002e-05,
-      "loss": 3.4508,
+      "epoch": 0.3707865168539326,
+      "grad_norm": 3.3499844074249268,
+      "learning_rate": 3.259259259259259e-05,
+      "loss": 0.805,
       "step": 66
     },
     {
-      "epoch": 0.3895553987297107,
-      "grad_norm": 11.202752113342285,
-      "learning_rate": 1.5681818181818182e-05,
-      "loss": 3.9702,
+      "epoch": 0.38764044943820225,
+      "grad_norm": 3.770142078399658,
+      "learning_rate": 3.4074074074074077e-05,
+      "loss": 0.7184,
       "step": 69
     },
     {
-      "epoch": 0.40649258997882853,
-      "grad_norm": 7.064818859100342,
-      "learning_rate": 1.6363636363636366e-05,
-      "loss": 3.5295,
+      "epoch": 0.4044943820224719,
+      "grad_norm": 3.740880012512207,
+      "learning_rate": 3.555555555555555e-05,
+      "loss": 0.8708,
       "step": 72
     },
     {
-      "epoch": 0.4234297812279464,
-      "grad_norm": 5.912719249725342,
-      "learning_rate": 1.7045454545454546e-05,
-      "loss": 3.6395,
+      "epoch": 0.42134831460674155,
+      "grad_norm": 2.981106996536255,
+      "learning_rate": 3.703703703703704e-05,
+      "loss": 0.8327,
       "step": 75
     },
     {
-      "epoch": 0.44036697247706424,
-      "grad_norm": 5.033207893371582,
-      "learning_rate": 1.772727272727273e-05,
-      "loss": 3.2398,
+      "epoch": 0.43820224719101125,
+      "grad_norm": 2.3469011783599854,
+      "learning_rate": 3.851851851851852e-05,
+      "loss": 0.5025,
       "step": 78
     },
     {
-      "epoch": 0.4573041637261821,
-      "grad_norm": 5.218384265899658,
-      "learning_rate": 1.840909090909091e-05,
-      "loss": 3.116,
+      "epoch": 0.4550561797752809,
+      "grad_norm": 3.296035051345825,
+      "learning_rate": 4e-05,
+      "loss": 0.6517,
       "step": 81
     },
     {
-      "epoch": 0.4573041637261821,
+      "epoch": 0.4550561797752809,
       "eval_NLI-v2_cosine_accuracy": 1.0,
       "eval_NLI-v2_dot_accuracy": 0.0,
       "eval_NLI-v2_euclidean_accuracy": 1.0,
       "eval_NLI-v2_manhattan_accuracy": 1.0,
       "eval_NLI-v2_max_accuracy": 1.0,
-      "eval_VitaminC_cosine_accuracy": 0.5546875,
-      "eval_VitaminC_cosine_accuracy_threshold": 0.9041332006454468,
-      "eval_VitaminC_cosine_ap": 0.5292859731465609,
-      "eval_VitaminC_cosine_f1": 0.6542553191489362,
-      "eval_VitaminC_cosine_f1_threshold": 0.452939510345459,
-      "eval_VitaminC_cosine_precision": 0.48616600790513836,
+      "eval_VitaminC_cosine_accuracy": 0.578125,
+      "eval_VitaminC_cosine_accuracy_threshold": 0.7859437465667725,
+      "eval_VitaminC_cosine_ap": 0.5557444337961499,
+      "eval_VitaminC_cosine_f1": 0.6595174262734584,
+      "eval_VitaminC_cosine_f1_threshold": 0.3211573362350464,
+      "eval_VitaminC_cosine_precision": 0.492,
       "eval_VitaminC_cosine_recall": 1.0,
-      "eval_VitaminC_dot_accuracy": 0.5546875,
-      "eval_VitaminC_dot_accuracy_threshold": 414.42559814453125,
-      "eval_VitaminC_dot_ap": 0.5222732504955002,
-      "eval_VitaminC_dot_f1": 0.6542553191489362,
-      "eval_VitaminC_dot_f1_threshold": 212.6934814453125,
-      "eval_VitaminC_dot_precision": 0.48616600790513836,
+      "eval_VitaminC_dot_accuracy": 0.578125,
+      "eval_VitaminC_dot_accuracy_threshold": 315.9444580078125,
+      "eval_VitaminC_dot_ap": 0.5539524528858992,
+      "eval_VitaminC_dot_f1": 0.6595174262734584,
+      "eval_VitaminC_dot_f1_threshold": 129.88558959960938,
+      "eval_VitaminC_dot_precision": 0.492,
       "eval_VitaminC_dot_recall": 1.0,
-      "eval_VitaminC_euclidean_accuracy": 0.5546875,
-      "eval_VitaminC_euclidean_accuracy_threshold": 9.18377685546875,
-      "eval_VitaminC_euclidean_ap": 0.5291787221346742,
-      "eval_VitaminC_euclidean_f1": 0.6542553191489362,
-      "eval_VitaminC_euclidean_f1_threshold": 22.683509826660156,
-      "eval_VitaminC_euclidean_precision": 0.48616600790513836,
+      "eval_VitaminC_euclidean_accuracy": 0.58203125,
+      "eval_VitaminC_euclidean_accuracy_threshold": 13.113249778747559,
+      "eval_VitaminC_euclidean_ap": 0.5510190217865811,
+      "eval_VitaminC_euclidean_f1": 0.6577540106951871,
+      "eval_VitaminC_euclidean_f1_threshold": 23.90462303161621,
+      "eval_VitaminC_euclidean_precision": 0.4900398406374502,
       "eval_VitaminC_euclidean_recall": 1.0,
-      "eval_VitaminC_manhattan_accuracy": 0.55859375,
-      "eval_VitaminC_manhattan_accuracy_threshold": 173.8212127685547,
-      "eval_VitaminC_manhattan_ap": 0.5305698453165033,
-      "eval_VitaminC_manhattan_f1": 0.6542553191489362,
-      "eval_VitaminC_manhattan_f1_threshold": 415.5366516113281,
-      "eval_VitaminC_manhattan_precision": 0.48616600790513836,
-      "eval_VitaminC_manhattan_recall": 1.0,
-      "eval_VitaminC_max_accuracy": 0.55859375,
-      "eval_VitaminC_max_accuracy_threshold": 414.42559814453125,
-      "eval_VitaminC_max_ap": 0.5305698453165033,
-      "eval_VitaminC_max_f1": 0.6542553191489362,
-      "eval_VitaminC_max_f1_threshold": 415.5366516113281,
-      "eval_VitaminC_max_precision": 0.48616600790513836,
+      "eval_VitaminC_manhattan_accuracy": 0.578125,
+      "eval_VitaminC_manhattan_accuracy_threshold": 276.40142822265625,
+      "eval_VitaminC_manhattan_ap": 0.5429240708188645,
+      "eval_VitaminC_manhattan_f1": 0.6576819407008085,
+      "eval_VitaminC_manhattan_f1_threshold": 469.7353515625,
+      "eval_VitaminC_manhattan_precision": 0.49193548387096775,
+      "eval_VitaminC_manhattan_recall": 0.991869918699187,
+      "eval_VitaminC_max_accuracy": 0.58203125,
+      "eval_VitaminC_max_accuracy_threshold": 315.9444580078125,
+      "eval_VitaminC_max_ap": 0.5557444337961499,
+      "eval_VitaminC_max_f1": 0.6595174262734584,
+      "eval_VitaminC_max_f1_threshold": 469.7353515625,
+      "eval_VitaminC_max_precision": 0.492,
       "eval_VitaminC_max_recall": 1.0,
-      "eval_sequential_score": 0.5305698453165033,
-      "eval_sts-test_pearson_cosine": 0.45494716382349193,
-      "eval_sts-test_pearson_dot": 0.44837123659858896,
-      "eval_sts-test_pearson_euclidean": 0.4480861256491879,
-      "eval_sts-test_pearson_manhattan": 0.4417008219313264,
-      "eval_sts-test_pearson_max": 0.45494716382349193,
-      "eval_sts-test_spearman_cosine": 0.48921418507251446,
-      "eval_sts-test_spearman_dot": 0.46707725062744593,
-      "eval_sts-test_spearman_euclidean": 0.4610824798409968,
-      "eval_sts-test_spearman_manhattan": 0.46068648052845956,
-      "eval_sts-test_spearman_max": 0.48921418507251446,
-      "eval_vitaminc-pairs_loss": 2.5043575763702393,
-      "eval_vitaminc-pairs_runtime": 1.4778,
-      "eval_vitaminc-pairs_samples_per_second": 73.079,
-      "eval_vitaminc-pairs_steps_per_second": 1.353,
+      "eval_sequential_score": 0.5557444337961499,
+      "eval_sts-test_pearson_cosine": 0.8483316632682467,
+      "eval_sts-test_pearson_dot": 0.8392403098680445,
+      "eval_sts-test_pearson_euclidean": 0.8814283057813619,
+      "eval_sts-test_pearson_manhattan": 0.8815226866327923,
+      "eval_sts-test_pearson_max": 0.8815226866327923,
+      "eval_sts-test_spearman_cosine": 0.8903503892346,
+      "eval_sts-test_spearman_dot": 0.857844431199042,
+      "eval_sts-test_spearman_euclidean": 0.8851830636663006,
+      "eval_sts-test_spearman_manhattan": 0.8865568876827619,
+      "eval_sts-test_spearman_max": 0.8903503892346,
+      "eval_vitaminc-pairs_loss": 2.3538782596588135,
+      "eval_vitaminc-pairs_runtime": 1.4618,
+      "eval_vitaminc-pairs_samples_per_second": 73.88,
+      "eval_vitaminc-pairs_steps_per_second": 1.368,
       "step": 81
     },
     {
-      "epoch": 0.4573041637261821,
-      "eval_negation-triplets_loss": 3.4229447841644287,
-      "eval_negation-triplets_runtime": 0.2991,
-      "eval_negation-triplets_samples_per_second": 213.954,
-      "eval_negation-triplets_steps_per_second": 3.343,
+      "epoch": 0.4550561797752809,
+      "eval_negation-triplets_loss": 1.649215579032898,
+      "eval_negation-triplets_runtime": 0.3081,
+      "eval_negation-triplets_samples_per_second": 207.723,
+      "eval_negation-triplets_steps_per_second": 3.246,
       "step": 81
     },
     {
-      "epoch": 0.4573041637261821,
-      "eval_scitail-pairs-pos_loss": 0.2784869372844696,
-      "eval_scitail-pairs-pos_runtime": 0.3633,
-      "eval_scitail-pairs-pos_samples_per_second": 148.649,
-      "eval_scitail-pairs-pos_steps_per_second": 2.753,
+      "epoch": 0.4550561797752809,
+      "eval_scitail-pairs-pos_loss": 0.11823470890522003,
+      "eval_scitail-pairs-pos_runtime": 0.376,
+      "eval_scitail-pairs-pos_samples_per_second": 143.616,
+      "eval_scitail-pairs-pos_steps_per_second": 2.66,
       "step": 81
     },
     {
-      "epoch": 0.4573041637261821,
-      "eval_xsum-pairs_loss": 2.428964614868164,
-      "eval_xsum-pairs_runtime": 3.1548,
-      "eval_xsum-pairs_samples_per_second": 40.573,
-      "eval_xsum-pairs_steps_per_second": 0.634,
+      "epoch": 0.4550561797752809,
+      "eval_xsum-pairs_loss": 0.08420603722333908,
+      "eval_xsum-pairs_runtime": 3.1576,
+      "eval_xsum-pairs_samples_per_second": 40.538,
+      "eval_xsum-pairs_steps_per_second": 0.633,
       "step": 81
     },
     {
-      "epoch": 0.4573041637261821,
-      "eval_sciq_pairs_loss": 0.15256048738956451,
-      "eval_sciq_pairs_runtime": 3.2432,
-      "eval_sciq_pairs_samples_per_second": 39.467,
-      "eval_sciq_pairs_steps_per_second": 0.617,
+      "epoch": 0.4550561797752809,
+      "eval_sciq_pairs_loss": 0.034781794995069504,
+      "eval_sciq_pairs_runtime": 3.2597,
+      "eval_sciq_pairs_samples_per_second": 39.267,
+      "eval_sciq_pairs_steps_per_second": 0.614,
       "step": 81
     },
     {
-      "epoch": 0.4573041637261821,
-      "eval_qasc_pairs_loss": 1.2902077436447144,
-      "eval_qasc_pairs_runtime": 0.6211,
-      "eval_qasc_pairs_samples_per_second": 206.085,
-      "eval_qasc_pairs_steps_per_second": 3.22,
+      "epoch": 0.4550561797752809,
+      "eval_qasc_pairs_loss": 0.10597346723079681,
+      "eval_qasc_pairs_runtime": 0.6245,
+      "eval_qasc_pairs_samples_per_second": 204.979,
+      "eval_qasc_pairs_steps_per_second": 3.203,
       "step": 81
     },
     {
-      "epoch": 0.4573041637261821,
-      "eval_openbookqa_pairs_loss": 2.4784862995147705,
-      "eval_openbookqa_pairs_runtime": 0.5758,
-      "eval_openbookqa_pairs_samples_per_second": 222.308,
-      "eval_openbookqa_pairs_steps_per_second": 3.474,
+      "epoch": 0.4550561797752809,
+      "eval_openbookqa_pairs_loss": 0.7160983681678772,
+      "eval_openbookqa_pairs_runtime": 0.5767,
+      "eval_openbookqa_pairs_samples_per_second": 221.961,
+      "eval_openbookqa_pairs_steps_per_second": 3.468,
       "step": 81
     },
     {
-      "epoch": 0.4573041637261821,
-      "eval_msmarco_pairs_loss": 2.967724084854126,
-      "eval_msmarco_pairs_runtime": 1.2944,
-      "eval_msmarco_pairs_samples_per_second": 98.885,
-      "eval_msmarco_pairs_steps_per_second": 1.545,
+      "epoch": 0.4550561797752809,
+      "eval_msmarco_pairs_loss": 0.3454173803329468,
+      "eval_msmarco_pairs_runtime": 1.2912,
+      "eval_msmarco_pairs_samples_per_second": 99.134,
+      "eval_msmarco_pairs_steps_per_second": 1.549,
       "step": 81
     },
     {
-      "epoch": 0.4573041637261821,
-      "eval_nq_pairs_loss": 3.358661413192749,
-      "eval_nq_pairs_runtime": 2.3827,
-      "eval_nq_pairs_samples_per_second": 53.722,
-      "eval_nq_pairs_steps_per_second": 0.839,
+      "epoch": 0.4550561797752809,
+      "eval_nq_pairs_loss": 0.4442503750324249,
+      "eval_nq_pairs_runtime": 2.3854,
+      "eval_nq_pairs_samples_per_second": 53.659,
+      "eval_nq_pairs_steps_per_second": 0.838,
       "step": 81
     },
     {
-      "epoch": 0.4573041637261821,
-      "eval_trivia_pairs_loss": 3.1391680240631104,
-      "eval_trivia_pairs_runtime": 4.4155,
-      "eval_trivia_pairs_samples_per_second": 28.989,
-      "eval_trivia_pairs_steps_per_second": 0.453,
+      "epoch": 0.4550561797752809,
+      "eval_trivia_pairs_loss": 0.9324482679367065,
+      "eval_trivia_pairs_runtime": 4.4251,
+      "eval_trivia_pairs_samples_per_second": 28.926,
+      "eval_trivia_pairs_steps_per_second": 0.452,
       "step": 81
     },
     {
-      "epoch": 0.4573041637261821,
-      "eval_gooaq_pairs_loss": 2.8774912357330322,
-      "eval_gooaq_pairs_runtime": 0.8746,
-      "eval_gooaq_pairs_samples_per_second": 146.346,
-      "eval_gooaq_pairs_steps_per_second": 2.287,
+      "epoch": 0.4550561797752809,
+      "eval_gooaq_pairs_loss": 0.6094165444374084,
+      "eval_gooaq_pairs_runtime": 0.8751,
+      "eval_gooaq_pairs_samples_per_second": 146.261,
+      "eval_gooaq_pairs_steps_per_second": 2.285,
       "step": 81
     },
     {
-      "epoch": 0.4573041637261821,
-      "eval_paws-pos_loss": 0.19754411280155182,
-      "eval_paws-pos_runtime": 0.684,
-      "eval_paws-pos_samples_per_second": 187.141,
-      "eval_paws-pos_steps_per_second": 2.924,
+      "epoch": 0.4550561797752809,
+      "eval_paws-pos_loss": 0.024421451613307,
+      "eval_paws-pos_runtime": 0.6865,
+      "eval_paws-pos_samples_per_second": 186.444,
+      "eval_paws-pos_steps_per_second": 2.913,
       "step": 81
     },
     {
-      "epoch": 0.47424135497529996,
-      "grad_norm": 5.149569988250732,
-      "learning_rate": 1.9090909090909094e-05,
-      "loss": 2.6049,
+      "epoch": 0.47191011235955055,
+      "grad_norm": 3.1395561695098877,
+      "learning_rate": 3.999675367909485e-05,
+      "loss": 0.5801,
       "step": 84
     },
     {
-      "epoch": 0.4911785462244178,
-      "grad_norm": 5.012928009033203,
-      "learning_rate": 1.9772727272727274e-05,
-      "loss": 2.7738,
+      "epoch": 0.4887640449438202,
+      "grad_norm": 2.7977917194366455,
+      "learning_rate": 3.998701612152597e-05,
+      "loss": 0.791,
       "step": 87
     },
     {
-      "epoch": 0.5081157374735357,
-      "grad_norm": 4.880725383758545,
-      "learning_rate": 2.0454545454545457e-05,
-      "loss": 2.5416,
+      "epoch": 0.5056179775280899,
+      "grad_norm": 2.3682048320770264,
+      "learning_rate": 3.997079154212493e-05,
+      "loss": 0.6042,
       "step": 90
     },
     {
-      "epoch": 0.5250529287226535,
-      "grad_norm": 5.618528366088867,
-      "learning_rate": 2.113636363636364e-05,
-      "loss": 2.3913,
+      "epoch": 0.5224719101123596,
+      "grad_norm": 2.843482255935669,
+      "learning_rate": 3.99480869635839e-05,
+      "loss": 0.7559,
       "step": 93
     },
     {
-      "epoch": 0.5419901199717714,
-      "grad_norm": 5.020515441894531,
-      "learning_rate": 2.1818181818181818e-05,
-      "loss": 2.3144,
+      "epoch": 0.5393258426966292,
+      "grad_norm": 2.7346785068511963,
+      "learning_rate": 3.9918912213415936e-05,
+      "loss": 0.6258,
       "step": 96
     },
     {
-      "epoch": 0.5589273112208892,
-      "grad_norm": 4.818451404571533,
-      "learning_rate": 2.25e-05,
-      "loss": 2.1857,
+      "epoch": 0.5561797752808989,
+      "grad_norm": 3.149007558822632,
+      "learning_rate": 3.9883279919701226e-05,
+      "loss": 0.8853,
       "step": 99
     },
     {
-      "epoch": 0.5758645024700071,
-      "grad_norm": 5.094771385192871,
-      "learning_rate": 2.3181818181818185e-05,
-      "loss": 1.8881,
+      "epoch": 0.5730337078651685,
+      "grad_norm": 3.3424761295318604,
+      "learning_rate": 3.9841205505621106e-05,
+      "loss": 0.5947,
       "step": 102
     },
     {
-      "epoch": 0.592801693719125,
-      "grad_norm": 3.795962333679199,
-      "learning_rate": 2.3863636363636365e-05,
-      "loss": 2.2699,
+      "epoch": 0.5898876404494382,
+      "grad_norm": 2.6377146244049072,
+      "learning_rate": 3.979270718278224e-05,
+      "loss": 0.644,
       "step": 105
     },
     {
-      "epoch": 0.6097388849682428,
-      "grad_norm": 4.46245813369751,
-      "learning_rate": 2.454545454545455e-05,
-      "loss": 2.1425,
+      "epoch": 0.6067415730337079,
+      "grad_norm": 1.3963145017623901,
+      "learning_rate": 3.973780594333386e-05,
+      "loss": 0.5682,
       "step": 108
     },
     {
-      "epoch": 0.6097388849682428,
+      "epoch": 0.6067415730337079,
       "eval_NLI-v2_cosine_accuracy": 1.0,
       "eval_NLI-v2_dot_accuracy": 0.0,
       "eval_NLI-v2_euclidean_accuracy": 1.0,
       "eval_NLI-v2_manhattan_accuracy": 1.0,
       "eval_NLI-v2_max_accuracy": 1.0,
-      "eval_VitaminC_cosine_accuracy": 0.5546875,
-      "eval_VitaminC_cosine_accuracy_threshold": 0.8830112218856812,
-      "eval_VitaminC_cosine_ap": 0.5302172957740995,
-      "eval_VitaminC_cosine_f1": 0.6558265582655827,
-      "eval_VitaminC_cosine_f1_threshold": 0.5253933668136597,
-      "eval_VitaminC_cosine_precision": 0.491869918699187,
-      "eval_VitaminC_cosine_recall": 0.983739837398374,
-      "eval_VitaminC_dot_accuracy": 0.5390625,
-      "eval_VitaminC_dot_accuracy_threshold": 427.5576171875,
-      "eval_VitaminC_dot_ap": 0.517120157327104,
-      "eval_VitaminC_dot_f1": 0.6542553191489362,
-      "eval_VitaminC_dot_f1_threshold": 175.80963134765625,
-      "eval_VitaminC_dot_precision": 0.48616600790513836,
+      "eval_VitaminC_cosine_accuracy": 0.58984375,
+      "eval_VitaminC_cosine_accuracy_threshold": 0.7784540057182312,
+      "eval_VitaminC_cosine_ap": 0.556890553952148,
+      "eval_VitaminC_cosine_f1": 0.6595174262734584,
+      "eval_VitaminC_cosine_f1_threshold": 0.3015836775302887,
+      "eval_VitaminC_cosine_precision": 0.492,
+      "eval_VitaminC_cosine_recall": 1.0,
+      "eval_VitaminC_dot_accuracy": 0.57421875,
+      "eval_VitaminC_dot_accuracy_threshold": 310.07818603515625,
+      "eval_VitaminC_dot_ap": 0.5486679382699982,
+      "eval_VitaminC_dot_f1": 0.6595174262734584,
+      "eval_VitaminC_dot_f1_threshold": 110.96945190429688,
+      "eval_VitaminC_dot_precision": 0.492,
       "eval_VitaminC_dot_recall": 1.0,
-      "eval_VitaminC_euclidean_accuracy": 0.5625,
-      "eval_VitaminC_euclidean_accuracy_threshold": 10.817148208618164,
-      "eval_VitaminC_euclidean_ap": 0.532255112376416,
-      "eval_VitaminC_euclidean_f1": 0.6558265582655827,
-      "eval_VitaminC_euclidean_f1_threshold": 21.10729217529297,
-      "eval_VitaminC_euclidean_precision": 0.491869918699187,
-      "eval_VitaminC_euclidean_recall": 0.983739837398374,
-      "eval_VitaminC_manhattan_accuracy": 0.5546875,
-      "eval_VitaminC_manhattan_accuracy_threshold": 224.70416259765625,
-      "eval_VitaminC_manhattan_ap": 0.5298930718604624,
-      "eval_VitaminC_manhattan_f1": 0.6558265582655827,
-      "eval_VitaminC_manhattan_f1_threshold": 415.3311767578125,
-      "eval_VitaminC_manhattan_precision": 0.491869918699187,
+      "eval_VitaminC_euclidean_accuracy": 0.578125,
+      "eval_VitaminC_euclidean_accuracy_threshold": 13.410951614379883,
+      "eval_VitaminC_euclidean_ap": 0.5551150763619972,
+      "eval_VitaminC_euclidean_f1": 0.6577540106951871,
+      "eval_VitaminC_euclidean_f1_threshold": 23.933565139770508,
+      "eval_VitaminC_euclidean_precision": 0.4900398406374502,
+      "eval_VitaminC_euclidean_recall": 1.0,
+      "eval_VitaminC_manhattan_accuracy": 0.58203125,
+      "eval_VitaminC_manhattan_accuracy_threshold": 236.87246704101562,
+      "eval_VitaminC_manhattan_ap": 0.5465417933692003,
+      "eval_VitaminC_manhattan_f1": 0.6576086956521738,
+      "eval_VitaminC_manhattan_f1_threshold": 479.8819580078125,
+      "eval_VitaminC_manhattan_precision": 0.49387755102040815,
       "eval_VitaminC_manhattan_recall": 0.983739837398374,
-      "eval_VitaminC_max_accuracy": 0.5625,
-      "eval_VitaminC_max_accuracy_threshold": 427.5576171875,
-      "eval_VitaminC_max_ap": 0.532255112376416,
-      "eval_VitaminC_max_f1": 0.6558265582655827,
-      "eval_VitaminC_max_f1_threshold": 415.3311767578125,
-      "eval_VitaminC_max_precision": 0.491869918699187,
+      "eval_VitaminC_max_accuracy": 0.58984375,
+      "eval_VitaminC_max_accuracy_threshold": 310.07818603515625,
+      "eval_VitaminC_max_ap": 0.556890553952148,
+      "eval_VitaminC_max_f1": 0.6595174262734584,
+      "eval_VitaminC_max_f1_threshold": 479.8819580078125,
+      "eval_VitaminC_max_precision": 0.49387755102040815,
       "eval_VitaminC_max_recall": 1.0,
-      "eval_sequential_score": 0.532255112376416,
-      "eval_sts-test_pearson_cosine": 0.755963151531783,
-      "eval_sts-test_pearson_dot": 0.7384823091540473,
-      "eval_sts-test_pearson_euclidean": 0.764089555623164,
-      "eval_sts-test_pearson_manhattan": 0.7670467479701304,
-      "eval_sts-test_pearson_max": 0.7670467479701304,
-      "eval_sts-test_spearman_cosine": 0.7806331583677342,
-      "eval_sts-test_spearman_dot": 0.7442842883778696,
-      "eval_sts-test_spearman_euclidean": 0.7674205303105437,
-      "eval_sts-test_spearman_manhattan": 0.7664974867050092,
-      "eval_sts-test_spearman_max": 0.7806331583677342,
-      "eval_vitaminc-pairs_loss": 2.721674919128418,
-      "eval_vitaminc-pairs_runtime": 1.4468,
-      "eval_vitaminc-pairs_samples_per_second": 74.65,
-      "eval_vitaminc-pairs_steps_per_second": 1.382,
+      "eval_sequential_score": 0.556890553952148,
+      "eval_sts-test_pearson_cosine": 0.8554011700695058,
+      "eval_sts-test_pearson_dot": 0.8545651085608208,
+      "eval_sts-test_pearson_euclidean": 0.8842988585732054,
+      "eval_sts-test_pearson_manhattan": 0.8850809337540164,
+      "eval_sts-test_pearson_max": 0.8850809337540164,
+      "eval_sts-test_spearman_cosine": 0.8959079853731212,
+      "eval_sts-test_spearman_dot": 0.876135947365041,
+      "eval_sts-test_spearman_euclidean": 0.8856381406339634,
+      "eval_sts-test_spearman_manhattan": 0.8868948834793577,
+      "eval_sts-test_spearman_max": 0.8959079853731212,
+      "eval_vitaminc-pairs_loss": 2.4271271228790283,
+      "eval_vitaminc-pairs_runtime": 1.4672,
+      "eval_vitaminc-pairs_samples_per_second": 73.61,
+      "eval_vitaminc-pairs_steps_per_second": 1.363,
       "step": 108
     },
     {
-      "epoch": 0.6097388849682428,
-      "eval_negation-triplets_loss": 2.338909387588501,
-      "eval_negation-triplets_runtime": 0.3017,
-      "eval_negation-triplets_samples_per_second": 212.101,
-      "eval_negation-triplets_steps_per_second": 3.314,
+      "epoch": 0.6067415730337079,
+      "eval_negation-triplets_loss": 1.6174229383468628,
+      "eval_negation-triplets_runtime": 0.3094,
+      "eval_negation-triplets_samples_per_second": 206.869,
+      "eval_negation-triplets_steps_per_second": 3.232,
       "step": 108
     },
     {
-      "epoch": 0.6097388849682428,
-      "eval_scitail-pairs-pos_loss": 0.23291125893592834,
-      "eval_scitail-pairs-pos_runtime": 0.3664,
-      "eval_scitail-pairs-pos_samples_per_second": 147.385,
-      "eval_scitail-pairs-pos_steps_per_second": 2.729,
+      "epoch": 0.6067415730337079,
+      "eval_scitail-pairs-pos_loss": 0.10586681962013245,
+      "eval_scitail-pairs-pos_runtime": 0.4307,
+      "eval_scitail-pairs-pos_samples_per_second": 125.374,
+      "eval_scitail-pairs-pos_steps_per_second": 2.322,
       "step": 108
     },
     {
-      "epoch": 0.6097388849682428,
-      "eval_xsum-pairs_loss": 1.2065516710281372,
-      "eval_xsum-pairs_runtime": 3.1488,
-      "eval_xsum-pairs_samples_per_second": 40.65,
-      "eval_xsum-pairs_steps_per_second": 0.635,
+      "epoch": 0.6067415730337079,
+      "eval_xsum-pairs_loss": 0.1041470319032669,
+      "eval_xsum-pairs_runtime": 3.2912,
+      "eval_xsum-pairs_samples_per_second": 38.891,
+      "eval_xsum-pairs_steps_per_second": 0.608,
       "step": 108
     },
     {
-      "epoch": 0.6097388849682428,
-      "eval_sciq_pairs_loss": 0.09487833082675934,
-      "eval_sciq_pairs_runtime": 3.2618,
-      "eval_sciq_pairs_samples_per_second": 39.242,
-      "eval_sciq_pairs_steps_per_second": 0.613,
+      "epoch": 0.6067415730337079,
+      "eval_sciq_pairs_loss": 0.03364330902695656,
+      "eval_sciq_pairs_runtime": 3.3617,
+      "eval_sciq_pairs_samples_per_second": 38.076,
+      "eval_sciq_pairs_steps_per_second": 0.595,
       "step": 108
     },
     {
-      "epoch": 0.6097388849682428,
-      "eval_qasc_pairs_loss": 0.8461999297142029,
-      "eval_qasc_pairs_runtime": 0.6246,
-      "eval_qasc_pairs_samples_per_second": 204.93,
-      "eval_qasc_pairs_steps_per_second": 3.202,
+      "epoch": 0.6067415730337079,
+      "eval_qasc_pairs_loss": 0.10827122628688812,
+      "eval_qasc_pairs_runtime": 0.6238,
+      "eval_qasc_pairs_samples_per_second": 205.206,
+      "eval_qasc_pairs_steps_per_second": 3.206,
       "step": 108
     },
     {
-      "epoch": 0.6097388849682428,
-      "eval_openbookqa_pairs_loss": 1.5739191770553589,
-      "eval_openbookqa_pairs_runtime": 0.5751,
-      "eval_openbookqa_pairs_samples_per_second": 222.568,
-      "eval_openbookqa_pairs_steps_per_second": 3.478,
+      "epoch": 0.6067415730337079,
+      "eval_openbookqa_pairs_loss": 0.7513518929481506,
+      "eval_openbookqa_pairs_runtime": 0.5882,
+      "eval_openbookqa_pairs_samples_per_second": 217.628,
+      "eval_openbookqa_pairs_steps_per_second": 3.4,
       "step": 108
     },
     {
-      "epoch": 0.6097388849682428,
-      "eval_msmarco_pairs_loss": 1.6446179151535034,
-      "eval_msmarco_pairs_runtime": 1.2828,
-      "eval_msmarco_pairs_samples_per_second": 99.784,
-      "eval_msmarco_pairs_steps_per_second": 1.559,
+      "epoch": 0.6067415730337079,
+      "eval_msmarco_pairs_loss": 0.3447520434856415,
+      "eval_msmarco_pairs_runtime": 1.2968,
+      "eval_msmarco_pairs_samples_per_second": 98.705,
+      "eval_msmarco_pairs_steps_per_second": 1.542,
       "step": 108
     },
     {
-      "epoch": 0.6097388849682428,
-      "eval_nq_pairs_loss": 2.364896535873413,
-      "eval_nq_pairs_runtime": 2.3802,
-      "eval_nq_pairs_samples_per_second": 53.777,
-      "eval_nq_pairs_steps_per_second": 0.84,
+      "epoch": 0.6067415730337079,
+      "eval_nq_pairs_loss": 0.4307234287261963,
+      "eval_nq_pairs_runtime": 2.4164,
+      "eval_nq_pairs_samples_per_second": 52.971,
+      "eval_nq_pairs_steps_per_second": 0.828,
       "step": 108
     },
     {
-      "epoch": 0.6097388849682428,
-      "eval_trivia_pairs_loss": 1.7080069780349731,
-      "eval_trivia_pairs_runtime": 4.4372,
-      "eval_trivia_pairs_samples_per_second": 28.847,
-      "eval_trivia_pairs_steps_per_second": 0.451,
+      "epoch": 0.6067415730337079,
+      "eval_trivia_pairs_loss": 0.9260168075561523,
+      "eval_trivia_pairs_runtime": 4.4644,
+      "eval_trivia_pairs_samples_per_second": 28.671,
+      "eval_trivia_pairs_steps_per_second": 0.448,
       "step": 108
     },
     {
-      "epoch": 0.6097388849682428,
-      "eval_gooaq_pairs_loss": 1.7924479246139526,
-      "eval_gooaq_pairs_runtime": 0.8761,
-      "eval_gooaq_pairs_samples_per_second": 146.094,
-      "eval_gooaq_pairs_steps_per_second": 2.283,
+      "epoch": 0.6067415730337079,
+      "eval_gooaq_pairs_loss": 0.6301646828651428,
+      "eval_gooaq_pairs_runtime": 0.9092,
+      "eval_gooaq_pairs_samples_per_second": 140.777,
+      "eval_gooaq_pairs_steps_per_second": 2.2,
       "step": 108
     },
     {
-      "epoch": 0.6097388849682428,
-      "eval_paws-pos_loss": 0.08000019192695618,
-      "eval_paws-pos_runtime": 0.6839,
-      "eval_paws-pos_samples_per_second": 187.168,
-      "eval_paws-pos_steps_per_second": 2.924,
+      "epoch": 0.6067415730337079,
+      "eval_paws-pos_loss": 0.024637963622808456,
+      "eval_paws-pos_runtime": 0.696,
+      "eval_paws-pos_samples_per_second": 183.9,
+      "eval_paws-pos_steps_per_second": 2.873,
       "step": 108
     },
     {
-      "epoch": 0.6266760762173607,
-      "grad_norm": 4.418070316314697,
-      "learning_rate": 2.5227272727272732e-05,
-      "loss": 2.1276,
+      "epoch": 0.6235955056179775,
+      "grad_norm": 3.0150091648101807,
+      "learning_rate": 3.9676525550881484e-05,
+      "loss": 0.5974,
       "step": 111
     },
     {
-      "epoch": 0.6436132674664785,
-      "grad_norm": 4.3495259284973145,
-      "learning_rate": 2.590909090909091e-05,
-      "loss": 1.7531,
+      "epoch": 0.6404494382022472,
+      "grad_norm": 2.7985854148864746,
+      "learning_rate": 3.9608892530200996e-05,
+      "loss": 0.649,
       "step": 114
     },
     {
-      "epoch": 0.6605504587155964,
-      "grad_norm": 4.294332027435303,
-      "learning_rate": 2.6590909090909093e-05,
-      "loss": 2.0179,
+      "epoch": 0.6573033707865169,
+      "grad_norm": 3.00435471534729,
+      "learning_rate": 3.953493615575757e-05,
+      "loss": 0.6966,
       "step": 117
     },
     {
-      "epoch": 0.6774876499647142,
-      "grad_norm": 3.4215610027313232,
-      "learning_rate": 2.7272727272727273e-05,
-      "loss": 1.5305,
+      "epoch": 0.6741573033707865,
+      "grad_norm": 3.276264190673828,
+      "learning_rate": 3.945468843903448e-05,
+      "loss": 0.542,
       "step": 120
     },
     {
-      "epoch": 0.6944248412138321,
-      "grad_norm": 4.37844181060791,
-      "learning_rate": 2.7954545454545457e-05,
-      "loss": 1.6925,
+      "epoch": 0.6910112359550562,
+      "grad_norm": 3.144037961959839,
+      "learning_rate": 3.936818411467709e-05,
+      "loss": 0.8583,
       "step": 123
     },
     {
-      "epoch": 0.7113620324629499,
-      "grad_norm": 4.019878387451172,
-      "learning_rate": 2.863636363636364e-05,
-      "loss": 1.5248,
+      "epoch": 0.7078651685393258,
+      "grad_norm": 2.498800754547119,
+      "learning_rate": 3.9275460625458294e-05,
+      "loss": 0.6416,
       "step": 126
     },
     {
-      "epoch": 0.7282992237120678,
-      "grad_norm": 4.662445068359375,
-      "learning_rate": 2.931818181818182e-05,
-      "loss": 1.523,
+      "epoch": 0.7247191011235955,
+      "grad_norm": 3.01488995552063,
+      "learning_rate": 3.917655810607162e-05,
+      "loss": 0.6273,
       "step": 129
     },
     {
-      "epoch": 0.7452364149611856,
-      "grad_norm": 4.6323161125183105,
-      "learning_rate": 3.0000000000000004e-05,
-      "loss": 1.5474,
+      "epoch": 0.7415730337078652,
+      "grad_norm": 3.439838171005249,
+      "learning_rate": 3.907151936575922e-05,
+      "loss": 0.8621,
       "step": 132
     },
     {
-      "epoch": 0.7621736062103035,
-      "grad_norm": 4.586575984954834,
-      "learning_rate": 3.068181818181819e-05,
-      "loss": 1.7221,
+      "epoch": 0.7584269662921348,
+      "grad_norm": 3.0246922969818115,
+      "learning_rate": 3.896038986978224e-05,
+      "loss": 0.7221,
       "step": 135
     },
     {
-      "epoch": 0.7621736062103035,
+      "epoch": 0.7584269662921348,
       "eval_NLI-v2_cosine_accuracy": 1.0,
       "eval_NLI-v2_dot_accuracy": 0.0,
       "eval_NLI-v2_euclidean_accuracy": 1.0,
       "eval_NLI-v2_manhattan_accuracy": 1.0,
       "eval_NLI-v2_max_accuracy": 1.0,
-      "eval_VitaminC_cosine_accuracy": 0.56640625,
-      "eval_VitaminC_cosine_accuracy_threshold": 0.8478574156761169,
-      "eval_VitaminC_cosine_ap": 0.5325579595957614,
-      "eval_VitaminC_cosine_f1": 0.6559999999999999,
-      "eval_VitaminC_cosine_f1_threshold": 0.35839784145355225,
-      "eval_VitaminC_cosine_precision": 0.4880952380952381,
+      "eval_VitaminC_cosine_accuracy": 0.57421875,
+      "eval_VitaminC_cosine_accuracy_threshold": 0.8462294936180115,
+      "eval_VitaminC_cosine_ap": 0.5610960529859609,
+      "eval_VitaminC_cosine_f1": 0.6595174262734584,
+      "eval_VitaminC_cosine_f1_threshold": 0.3176865577697754,
+      "eval_VitaminC_cosine_precision": 0.492,
       "eval_VitaminC_cosine_recall": 1.0,
-      "eval_VitaminC_dot_accuracy": 0.5625,
-      "eval_VitaminC_dot_accuracy_threshold": 366.9839172363281,
-      "eval_VitaminC_dot_ap": 0.5326813797607027,
-      "eval_VitaminC_dot_f1": 0.6559999999999999,
-      "eval_VitaminC_dot_f1_threshold": 157.35829162597656,
-      "eval_VitaminC_dot_precision": 0.4880952380952381,
+      "eval_VitaminC_dot_accuracy": 0.5859375,
+      "eval_VitaminC_dot_accuracy_threshold": 319.20416259765625,
+      "eval_VitaminC_dot_ap": 0.5540963620104548,
+      "eval_VitaminC_dot_f1": 0.6595174262734584,
+      "eval_VitaminC_dot_f1_threshold": 121.18156433105469,
+      "eval_VitaminC_dot_precision": 0.492,
       "eval_VitaminC_dot_recall": 1.0,
-      "eval_VitaminC_euclidean_accuracy": 0.5625,
-      "eval_VitaminC_euclidean_accuracy_threshold": 12.044445037841797,
-      "eval_VitaminC_euclidean_ap": 0.5304103559932005,
-      "eval_VitaminC_euclidean_f1": 0.6542553191489362,
-      "eval_VitaminC_euclidean_f1_threshold": 24.461441040039062,
-      "eval_VitaminC_euclidean_precision": 0.48616600790513836,
+      "eval_VitaminC_euclidean_accuracy": 0.58203125,
+      "eval_VitaminC_euclidean_accuracy_threshold": 15.303094863891602,
+      "eval_VitaminC_euclidean_ap": 0.5590109366458975,
+      "eval_VitaminC_euclidean_f1": 0.6577540106951871,
+      "eval_VitaminC_euclidean_f1_threshold": 23.605255126953125,
+      "eval_VitaminC_euclidean_precision": 0.4900398406374502,
       "eval_VitaminC_euclidean_recall": 1.0,
-      "eval_VitaminC_manhattan_accuracy": 0.5625,
-      "eval_VitaminC_manhattan_accuracy_threshold": 239.24815368652344,
-      "eval_VitaminC_manhattan_ap": 0.5314780667834758,
-      "eval_VitaminC_manhattan_f1": 0.6575342465753424,
-      "eval_VitaminC_manhattan_f1_threshold": 400.6834716796875,
-      "eval_VitaminC_manhattan_precision": 0.49586776859504134,
-      "eval_VitaminC_manhattan_recall": 0.975609756097561,
-      "eval_VitaminC_max_accuracy": 0.56640625,
-      "eval_VitaminC_max_accuracy_threshold": 366.9839172363281,
-      "eval_VitaminC_max_ap": 0.5326813797607027,
-      "eval_VitaminC_max_f1": 0.6575342465753424,
-      "eval_VitaminC_max_f1_threshold": 400.6834716796875,
-      "eval_VitaminC_max_precision": 0.49586776859504134,
+      "eval_VitaminC_manhattan_accuracy": 0.578125,
+      "eval_VitaminC_manhattan_accuracy_threshold": 240.15660095214844,
+      "eval_VitaminC_manhattan_ap": 0.5523438512618317,
+      "eval_VitaminC_manhattan_f1": 0.6577540106951871,
+      "eval_VitaminC_manhattan_f1_threshold": 503.4666748046875,
+      "eval_VitaminC_manhattan_precision": 0.4900398406374502,
+      "eval_VitaminC_manhattan_recall": 1.0,
+      "eval_VitaminC_max_accuracy": 0.5859375,
+      "eval_VitaminC_max_accuracy_threshold": 319.20416259765625,
+      "eval_VitaminC_max_ap": 0.5610960529859609,
+      "eval_VitaminC_max_f1": 0.6595174262734584,
+      "eval_VitaminC_max_f1_threshold": 503.4666748046875,
+      "eval_VitaminC_max_precision": 0.492,
       "eval_VitaminC_max_recall": 1.0,
-      "eval_sequential_score": 0.5326813797607027,
-      "eval_sts-test_pearson_cosine": 0.7919597804368175,
-      "eval_sts-test_pearson_dot": 0.7994867531185785,
-      "eval_sts-test_pearson_euclidean": 0.8117960113303863,
-      "eval_sts-test_pearson_manhattan": 0.8144714466358016,
-      "eval_sts-test_pearson_max": 0.8144714466358016,
-      "eval_sts-test_spearman_cosine": 0.831478610786181,
-      "eval_sts-test_spearman_dot": 0.8192534746855707,
-      "eval_sts-test_spearman_euclidean": 0.8185577905406703,
-      "eval_sts-test_spearman_manhattan": 0.8154771593606782,
-      "eval_sts-test_spearman_max": 0.831478610786181,
-      "eval_vitaminc-pairs_loss": 2.852091073989868,
-      "eval_vitaminc-pairs_runtime": 1.4427,
-      "eval_vitaminc-pairs_samples_per_second": 74.858,
-      "eval_vitaminc-pairs_steps_per_second": 1.386,
+      "eval_sequential_score": 0.5610960529859609,
+      "eval_sts-test_pearson_cosine": 0.8558158452024208,
+      "eval_sts-test_pearson_dot": 0.8576514422982167,
+      "eval_sts-test_pearson_euclidean": 0.8865043101846597,
+      "eval_sts-test_pearson_manhattan": 0.8866527906896583,
+      "eval_sts-test_pearson_max": 0.8866527906896583,
+      "eval_sts-test_spearman_cosine": 0.8964919590428757,
+      "eval_sts-test_spearman_dot": 0.8785377225806833,
+      "eval_sts-test_spearman_euclidean": 0.8874999818863996,
+      "eval_sts-test_spearman_manhattan": 0.8881272791699125,
+      "eval_sts-test_spearman_max": 0.8964919590428757,
+      "eval_vitaminc-pairs_loss": 2.336690902709961,
+      "eval_vitaminc-pairs_runtime": 1.4561,
+      "eval_vitaminc-pairs_samples_per_second": 74.172,
+      "eval_vitaminc-pairs_steps_per_second": 1.374,
       "step": 135
     },
     {
-      "epoch": 0.7621736062103035,
-      "eval_negation-triplets_loss": 2.074247121810913,
-      "eval_negation-triplets_runtime": 0.3,
-      "eval_negation-triplets_samples_per_second": 213.353,
-      "eval_negation-triplets_steps_per_second": 3.334,
+      "epoch": 0.7584269662921348,
+      "eval_negation-triplets_loss": 1.5712968111038208,
+      "eval_negation-triplets_runtime": 0.2973,
+      "eval_negation-triplets_samples_per_second": 215.304,
+      "eval_negation-triplets_steps_per_second": 3.364,
       "step": 135
     },
     {
-      "epoch": 0.7621736062103035,
-      "eval_scitail-pairs-pos_loss": 0.2149849385023117,
-      "eval_scitail-pairs-pos_runtime": 0.3744,
-      "eval_scitail-pairs-pos_samples_per_second": 144.219,
-      "eval_scitail-pairs-pos_steps_per_second": 2.671,
+      "epoch": 0.7584269662921348,
+      "eval_scitail-pairs-pos_loss": 0.1093834936618805,
+      "eval_scitail-pairs-pos_runtime": 0.3709,
+      "eval_scitail-pairs-pos_samples_per_second": 145.598,
+      "eval_scitail-pairs-pos_steps_per_second": 2.696,
       "step": 135
     },
     {
-      "epoch": 0.7621736062103035,
-      "eval_xsum-pairs_loss": 0.7706837058067322,
-      "eval_xsum-pairs_runtime": 3.1609,
-      "eval_xsum-pairs_samples_per_second": 40.495,
-      "eval_xsum-pairs_steps_per_second": 0.633,
+      "epoch": 0.7584269662921348,
+      "eval_xsum-pairs_loss": 0.09297582507133484,
+      "eval_xsum-pairs_runtime": 3.1496,
+      "eval_xsum-pairs_samples_per_second": 40.64,
+      "eval_xsum-pairs_steps_per_second": 0.635,
       "step": 135
     },
     {
-      "epoch": 0.7621736062103035,
-      "eval_sciq_pairs_loss": 0.07513368874788284,
-      "eval_sciq_pairs_runtime": 3.2949,
-      "eval_sciq_pairs_samples_per_second": 38.848,
-      "eval_sciq_pairs_steps_per_second": 0.607,
+      "epoch": 0.7584269662921348,
+      "eval_sciq_pairs_loss": 0.031587302684783936,
+      "eval_sciq_pairs_runtime": 3.2666,
+      "eval_sciq_pairs_samples_per_second": 39.185,
+      "eval_sciq_pairs_steps_per_second": 0.612,
       "step": 135
     },
     {
-      "epoch": 0.7621736062103035,
-      "eval_qasc_pairs_loss": 0.6355602741241455,
-      "eval_qasc_pairs_runtime": 0.6392,
-      "eval_qasc_pairs_samples_per_second": 200.246,
-      "eval_qasc_pairs_steps_per_second": 3.129,
+      "epoch": 0.7584269662921348,
+      "eval_qasc_pairs_loss": 0.10210572183132172,
+      "eval_qasc_pairs_runtime": 0.6258,
+      "eval_qasc_pairs_samples_per_second": 204.538,
+      "eval_qasc_pairs_steps_per_second": 3.196,
       "step": 135
     },
     {
-      "epoch": 0.7621736062103035,
-      "eval_openbookqa_pairs_loss": 1.4014525413513184,
-      "eval_openbookqa_pairs_runtime": 0.622,
-      "eval_openbookqa_pairs_samples_per_second": 205.786,
-      "eval_openbookqa_pairs_steps_per_second": 3.215,
+      "epoch": 0.7584269662921348,
+      "eval_openbookqa_pairs_loss": 0.7485109567642212,
+      "eval_openbookqa_pairs_runtime": 0.5836,
+      "eval_openbookqa_pairs_samples_per_second": 219.317,
+      "eval_openbookqa_pairs_steps_per_second": 3.427,
       "step": 135
     },
     {
-      "epoch": 0.7621736062103035,
-      "eval_msmarco_pairs_loss": 1.1524099111557007,
-      "eval_msmarco_pairs_runtime": 1.31,
-      "eval_msmarco_pairs_samples_per_second": 97.709,
-      "eval_msmarco_pairs_steps_per_second": 1.527,
+      "epoch": 0.7584269662921348,
+      "eval_msmarco_pairs_loss": 0.34248754382133484,
+      "eval_msmarco_pairs_runtime": 1.286,
+      "eval_msmarco_pairs_samples_per_second": 99.53,
+      "eval_msmarco_pairs_steps_per_second": 1.555,
       "step": 135
     },
     {
-      "epoch": 0.7621736062103035,
-      "eval_nq_pairs_loss": 1.7768574953079224,
-      "eval_nq_pairs_runtime": 2.3979,
-      "eval_nq_pairs_samples_per_second": 53.379,
-      "eval_nq_pairs_steps_per_second": 0.834,
+      "epoch": 0.7584269662921348,
+      "eval_nq_pairs_loss": 0.41257673501968384,
+      "eval_nq_pairs_runtime": 2.3755,
+      "eval_nq_pairs_samples_per_second": 53.883,
+      "eval_nq_pairs_steps_per_second": 0.842,
       "step": 135
     },
     {
-      "epoch": 0.7621736062103035,
-      "eval_trivia_pairs_loss": 1.4495295286178589,
-      "eval_trivia_pairs_runtime": 4.4194,
-      "eval_trivia_pairs_samples_per_second": 28.964,
+      "epoch": 0.7584269662921348,
+      "eval_trivia_pairs_loss": 0.9274640083312988,
+      "eval_trivia_pairs_runtime": 4.4185,
+      "eval_trivia_pairs_samples_per_second": 28.969,
       "eval_trivia_pairs_steps_per_second": 0.453,
       "step": 135
     },
     {
-      "epoch": 0.7621736062103035,
-      "eval_gooaq_pairs_loss": 1.3955378532409668,
-      "eval_gooaq_pairs_runtime": 0.8788,
-      "eval_gooaq_pairs_samples_per_second": 145.649,
-      "eval_gooaq_pairs_steps_per_second": 2.276,
+      "epoch": 0.7584269662921348,
+      "eval_gooaq_pairs_loss": 0.5840359330177307,
+      "eval_gooaq_pairs_runtime": 0.8784,
+      "eval_gooaq_pairs_samples_per_second": 145.726,
+      "eval_gooaq_pairs_steps_per_second": 2.277,
       "step": 135
     },
     {
-      "epoch": 0.7621736062103035,
-      "eval_paws-pos_loss": 0.06006813049316406,
-      "eval_paws-pos_runtime": 0.6896,
-      "eval_paws-pos_samples_per_second": 185.603,
-      "eval_paws-pos_steps_per_second": 2.9,
+      "epoch": 0.7584269662921348,
+      "eval_paws-pos_loss": 0.024575484916567802,
+      "eval_paws-pos_runtime": 0.6869,
+      "eval_paws-pos_samples_per_second": 186.355,
+      "eval_paws-pos_steps_per_second": 2.912,
       "step": 135
     },
     {
-      "epoch": 0.7791107974594214,
-      "grad_norm": 3.864208936691284,
-      "learning_rate": 3.1363636363636365e-05,
-      "loss": 1.5366,
+      "epoch": 0.7752808988764045,
+      "grad_norm": 3.189115285873413,
+      "learning_rate": 3.884321771974146e-05,
+      "loss": 0.9421,
       "step": 138
     },
     {
-      "epoch": 0.7960479887085392,
-      "grad_norm": 3.837550640106201,
-      "learning_rate": 3.204545454545455e-05,
-      "loss": 1.3045,
+      "epoch": 0.7921348314606742,
+      "grad_norm": 1.8421012163162231,
+      "learning_rate": 3.872005363275693e-05,
+      "loss": 0.6845,
       "step": 141
     },
     {
-      "epoch": 0.8129851799576571,
-      "grad_norm": 3.5258102416992188,
-      "learning_rate": 3.272727272727273e-05,
-      "loss": 1.1999,
+      "epoch": 0.8089887640449438,
+      "grad_norm": 2.2710273265838623,
+      "learning_rate": 3.859095091951534e-05,
+      "loss": 0.5464,
       "step": 144
     },
     {
-      "epoch": 0.8299223712067749,
-      "grad_norm": 3.4431183338165283,
-      "learning_rate": 3.340909090909091e-05,
-      "loss": 1.3483,
+      "epoch": 0.8258426966292135,
+      "grad_norm": 2.98201060295105,
+      "learning_rate": 3.845596546119496e-05,
+      "loss": 0.6338,
       "step": 147
     },
     {
-      "epoch": 0.8468595624558928,
-      "grad_norm": 3.6455864906311035,
-      "learning_rate": 3.409090909090909e-05,
-      "loss": 1.2009,
+      "epoch": 0.8426966292134831,
+      "grad_norm": 2.9222068786621094,
+      "learning_rate": 3.831515568527782e-05,
+      "loss": 0.4993,
       "step": 150
     },
     {
-      "epoch": 0.8637967537050106,
-      "grad_norm": 4.508525371551514,
-      "learning_rate": 3.4772727272727276e-05,
-      "loss": 1.4495,
+      "epoch": 0.8595505617977528,
+      "grad_norm": 3.6596696376800537,
+      "learning_rate": 3.81685825402598e-05,
+      "loss": 0.6939,
       "step": 153
     },
     {
-      "epoch": 0.8807339449541285,
-      "grad_norm": 3.0432400703430176,
-      "learning_rate": 3.545454545454546e-05,
-      "loss": 1.2329,
+      "epoch": 0.8764044943820225,
+      "grad_norm": 3.5363073348999023,
+      "learning_rate": 3.801630946926956e-05,
+      "loss": 0.5791,
       "step": 156
     },
     {
-      "epoch": 0.8976711362032463,
-      "grad_norm": 3.0190365314483643,
-      "learning_rate": 3.613636363636364e-05,
-      "loss": 1.1905,
+      "epoch": 0.8932584269662921,
+      "grad_norm": 3.6599326133728027,
+      "learning_rate": 3.785840238260758e-05,
+      "loss": 0.9226,
       "step": 159
     },
     {
-      "epoch": 0.9146083274523642,
-      "grad_norm": 3.74668288230896,
-      "learning_rate": 3.681818181818182e-05,
-      "loss": 1.277,
+      "epoch": 0.9101123595505618,
+      "grad_norm": 2.6967382431030273,
+      "learning_rate": 3.7694929629217385e-05,
+      "loss": 0.6336,
       "step": 162
     },
     {
-      "epoch": 0.9146083274523642,
+      "epoch": 0.9101123595505618,
       "eval_NLI-v2_cosine_accuracy": 1.0,
       "eval_NLI-v2_dot_accuracy": 0.0,
       "eval_NLI-v2_euclidean_accuracy": 1.0,
       "eval_NLI-v2_manhattan_accuracy": 1.0,
       "eval_NLI-v2_max_accuracy": 1.0,
-      "eval_VitaminC_cosine_accuracy": 0.57421875,
-      "eval_VitaminC_cosine_accuracy_threshold": 0.8101799488067627,
-      "eval_VitaminC_cosine_ap": 0.5298515171639175,
-      "eval_VitaminC_cosine_f1": 0.6542553191489362,
-      "eval_VitaminC_cosine_f1_threshold": 0.345889687538147,
-      "eval_VitaminC_cosine_precision": 0.48616600790513836,
+      "eval_VitaminC_cosine_accuracy": 0.578125,
+      "eval_VitaminC_cosine_accuracy_threshold": 0.8077100515365601,
+      "eval_VitaminC_cosine_ap": 0.560345569715395,
+      "eval_VitaminC_cosine_f1": 0.6577540106951871,
+      "eval_VitaminC_cosine_f1_threshold": 0.25425243377685547,
+      "eval_VitaminC_cosine_precision": 0.4900398406374502,
       "eval_VitaminC_cosine_recall": 1.0,
-      "eval_VitaminC_dot_accuracy": 0.55078125,
-      "eval_VitaminC_dot_accuracy_threshold": 373.5804443359375,
-      "eval_VitaminC_dot_ap": 0.5310954683437364,
-      "eval_VitaminC_dot_f1": 0.6542553191489362,
-      "eval_VitaminC_dot_f1_threshold": 155.41326904296875,
-      "eval_VitaminC_dot_precision": 0.48616600790513836,
-      "eval_VitaminC_dot_recall": 1.0,
-      "eval_VitaminC_euclidean_accuracy": 0.57421875,
-      "eval_VitaminC_euclidean_accuracy_threshold": 13.60124683380127,
-      "eval_VitaminC_euclidean_ap": 0.5286057955992807,
-      "eval_VitaminC_euclidean_f1": 0.6577540106951871,
-      "eval_VitaminC_euclidean_f1_threshold": 22.904512405395508,
-      "eval_VitaminC_euclidean_precision": 0.4900398406374502,
+      "eval_VitaminC_dot_accuracy": 0.578125,
+      "eval_VitaminC_dot_accuracy_threshold": 331.0595703125,
+      "eval_VitaminC_dot_ap": 0.5499174718252662,
+      "eval_VitaminC_dot_f1": 0.6594594594594595,
+      "eval_VitaminC_dot_f1_threshold": 121.5512924194336,
+      "eval_VitaminC_dot_precision": 0.4939271255060729,
+      "eval_VitaminC_dot_recall": 0.991869918699187,
+      "eval_VitaminC_euclidean_accuracy": 0.58203125,
+      "eval_VitaminC_euclidean_accuracy_threshold": 13.087348937988281,
+      "eval_VitaminC_euclidean_ap": 0.5563988051869968,
+      "eval_VitaminC_euclidean_f1": 0.6595174262734584,
+      "eval_VitaminC_euclidean_f1_threshold": 24.085674285888672,
+      "eval_VitaminC_euclidean_precision": 0.492,
       "eval_VitaminC_euclidean_recall": 1.0,
-      "eval_VitaminC_manhattan_accuracy": 0.57421875,
-      "eval_VitaminC_manhattan_accuracy_threshold": 262.37322998046875,
-      "eval_VitaminC_manhattan_ap": 0.5253560845853567,
-      "eval_VitaminC_manhattan_f1": 0.6559999999999999,
-      "eval_VitaminC_manhattan_f1_threshold": 465.94549560546875,
-      "eval_VitaminC_manhattan_precision": 0.4880952380952381,
+      "eval_VitaminC_manhattan_accuracy": 0.58203125,
+      "eval_VitaminC_manhattan_accuracy_threshold": 339.80157470703125,
+      "eval_VitaminC_manhattan_ap": 0.5563552623581395,
+      "eval_VitaminC_manhattan_f1": 0.6595174262734584,
+      "eval_VitaminC_manhattan_f1_threshold": 513.6148681640625,
+      "eval_VitaminC_manhattan_precision": 0.492,
       "eval_VitaminC_manhattan_recall": 1.0,
-      "eval_VitaminC_max_accuracy": 0.57421875,
-      "eval_VitaminC_max_accuracy_threshold": 373.5804443359375,
-      "eval_VitaminC_max_ap": 0.5310954683437364,
-      "eval_VitaminC_max_f1": 0.6577540106951871,
-      "eval_VitaminC_max_f1_threshold": 465.94549560546875,
-      "eval_VitaminC_max_precision": 0.4900398406374502,
+      "eval_VitaminC_max_accuracy": 0.58203125,
+      "eval_VitaminC_max_accuracy_threshold": 339.80157470703125,
+      "eval_VitaminC_max_ap": 0.560345569715395,
+      "eval_VitaminC_max_f1": 0.6595174262734584,
+      "eval_VitaminC_max_f1_threshold": 513.6148681640625,
+      "eval_VitaminC_max_precision": 0.4939271255060729,
       "eval_VitaminC_max_recall": 1.0,
-      "eval_sequential_score": 0.5310954683437364,
-      "eval_sts-test_pearson_cosine": 0.8067612938723231,
-      "eval_sts-test_pearson_dot": 0.8217874837658639,
-      "eval_sts-test_pearson_euclidean": 0.827948115812785,
-      "eval_sts-test_pearson_manhattan": 0.8261527694953693,
-      "eval_sts-test_pearson_max": 0.827948115812785,
-      "eval_sts-test_spearman_cosine": 0.8547777638284432,
-      "eval_sts-test_spearman_dot": 0.8498786150097738,
-      "eval_sts-test_spearman_euclidean": 0.8373845860667446,
-      "eval_sts-test_spearman_manhattan": 0.8324507067477893,
-      "eval_sts-test_spearman_max": 0.8547777638284432,
-      "eval_vitaminc-pairs_loss": 2.776399612426758,
-      "eval_vitaminc-pairs_runtime": 1.4503,
-      "eval_vitaminc-pairs_samples_per_second": 74.467,
-      "eval_vitaminc-pairs_steps_per_second": 1.379,
+      "eval_sequential_score": 0.560345569715395,
+      "eval_sts-test_pearson_cosine": 0.8572374144718338,
+      "eval_sts-test_pearson_dot": 0.8574922304214377,
+      "eval_sts-test_pearson_euclidean": 0.8855365860444931,
+      "eval_sts-test_pearson_manhattan": 0.886788554614621,
+      "eval_sts-test_pearson_max": 0.886788554614621,
+      "eval_sts-test_spearman_cosine": 0.8965259734503722,
+      "eval_sts-test_spearman_dot": 0.8784832995286892,
+      "eval_sts-test_spearman_euclidean": 0.8855371715506584,
+      "eval_sts-test_spearman_manhattan": 0.8887295132057965,
+      "eval_sts-test_spearman_max": 0.8965259734503722,
+      "eval_vitaminc-pairs_loss": 2.376066207885742,
+      "eval_vitaminc-pairs_runtime": 1.4378,
+      "eval_vitaminc-pairs_samples_per_second": 75.117,
+      "eval_vitaminc-pairs_steps_per_second": 1.391,
       "step": 162
     },
     {
-      "epoch": 0.9146083274523642,
-      "eval_negation-triplets_loss": 2.005451202392578,
-      "eval_negation-triplets_runtime": 0.2981,
-      "eval_negation-triplets_samples_per_second": 214.709,
-      "eval_negation-triplets_steps_per_second": 3.355,
+      "epoch": 0.9101123595505618,
+      "eval_negation-triplets_loss": 1.63050377368927,
+      "eval_negation-triplets_runtime": 0.3004,
+      "eval_negation-triplets_samples_per_second": 213.075,
+      "eval_negation-triplets_steps_per_second": 3.329,
       "step": 162
     },
     {
-      "epoch": 0.9146083274523642,
-      "eval_scitail-pairs-pos_loss": 0.19877880811691284,
-      "eval_scitail-pairs-pos_runtime": 0.3623,
-      "eval_scitail-pairs-pos_samples_per_second": 149.043,
-      "eval_scitail-pairs-pos_steps_per_second": 2.76,
+      "epoch": 0.9101123595505618,
+      "eval_scitail-pairs-pos_loss": 0.10006655752658844,
+      "eval_scitail-pairs-pos_runtime": 0.365,
+      "eval_scitail-pairs-pos_samples_per_second": 147.952,
+      "eval_scitail-pairs-pos_steps_per_second": 2.74,
       "step": 162
     },
     {
-      "epoch": 0.9146083274523642,
-      "eval_xsum-pairs_loss": 0.5586928725242615,
-      "eval_xsum-pairs_runtime": 3.1466,
-      "eval_xsum-pairs_samples_per_second": 40.679,
-      "eval_xsum-pairs_steps_per_second": 0.636,
+      "epoch": 0.9101123595505618,
+      "eval_xsum-pairs_loss": 0.0762382224202156,
+      "eval_xsum-pairs_runtime": 3.1586,
+      "eval_xsum-pairs_samples_per_second": 40.525,
+      "eval_xsum-pairs_steps_per_second": 0.633,
       "step": 162
     },
     {
-      "epoch": 0.9146083274523642,
-      "eval_sciq_pairs_loss": 0.06038254499435425,
-      "eval_sciq_pairs_runtime": 3.4092,
-      "eval_sciq_pairs_samples_per_second": 37.545,
-      "eval_sciq_pairs_steps_per_second": 0.587,
+      "epoch": 0.9101123595505618,
+      "eval_sciq_pairs_loss": 0.032084282487630844,
+      "eval_sciq_pairs_runtime": 3.2586,
+      "eval_sciq_pairs_samples_per_second": 39.28,
+      "eval_sciq_pairs_steps_per_second": 0.614,
       "step": 162
     },
     {
-      "epoch": 0.9146083274523642,
-      "eval_qasc_pairs_loss": 0.49434012174606323,
-      "eval_qasc_pairs_runtime": 0.6342,
-      "eval_qasc_pairs_samples_per_second": 201.832,
-      "eval_qasc_pairs_steps_per_second": 3.154,
+      "epoch": 0.9101123595505618,
+      "eval_qasc_pairs_loss": 0.10310036689043045,
+      "eval_qasc_pairs_runtime": 0.6197,
+      "eval_qasc_pairs_samples_per_second": 206.545,
+      "eval_qasc_pairs_steps_per_second": 3.227,
       "step": 162
     },
     {
-      "epoch": 0.9146083274523642,
-      "eval_openbookqa_pairs_loss": 1.1903400421142578,
-      "eval_openbookqa_pairs_runtime": 0.5754,
-      "eval_openbookqa_pairs_samples_per_second": 222.449,
-      "eval_openbookqa_pairs_steps_per_second": 3.476,
+      "epoch": 0.9101123595505618,
+      "eval_openbookqa_pairs_loss": 0.6995278000831604,
+      "eval_openbookqa_pairs_runtime": 0.575,
+      "eval_openbookqa_pairs_samples_per_second": 222.624,
+      "eval_openbookqa_pairs_steps_per_second": 3.478,
       "step": 162
     },
     {
-      "epoch": 0.9146083274523642,
-      "eval_msmarco_pairs_loss": 0.8656420707702637,
-      "eval_msmarco_pairs_runtime": 1.2858,
-      "eval_msmarco_pairs_samples_per_second": 99.547,
+      "epoch": 0.9101123595505618,
+      "eval_msmarco_pairs_loss": 0.37089064717292786,
+      "eval_msmarco_pairs_runtime": 1.2866,
+      "eval_msmarco_pairs_samples_per_second": 99.49,
       "eval_msmarco_pairs_steps_per_second": 1.555,
       "step": 162
     },
     {
-      "epoch": 0.9146083274523642,
-      "eval_nq_pairs_loss": 1.1553651094436646,
-      "eval_nq_pairs_runtime": 2.3754,
-      "eval_nq_pairs_samples_per_second": 53.885,
-      "eval_nq_pairs_steps_per_second": 0.842,
+      "epoch": 0.9101123595505618,
+      "eval_nq_pairs_loss": 0.39078566431999207,
+      "eval_nq_pairs_runtime": 2.3928,
+      "eval_nq_pairs_samples_per_second": 53.493,
+      "eval_nq_pairs_steps_per_second": 0.836,
       "step": 162
     },
     {
-      "epoch": 0.9146083274523642,
-      "eval_trivia_pairs_loss": 1.2928619384765625,
-      "eval_trivia_pairs_runtime": 4.4084,
-      "eval_trivia_pairs_samples_per_second": 29.035,
-      "eval_trivia_pairs_steps_per_second": 0.454,
+      "epoch": 0.9101123595505618,
+      "eval_trivia_pairs_loss": 0.9003691077232361,
+      "eval_trivia_pairs_runtime": 4.479,
+      "eval_trivia_pairs_samples_per_second": 28.578,
+      "eval_trivia_pairs_steps_per_second": 0.447,
       "step": 162
     },
     {
-      "epoch": 0.9146083274523642,
-      "eval_gooaq_pairs_loss": 1.1580811738967896,
-      "eval_gooaq_pairs_runtime": 0.8731,
-      "eval_gooaq_pairs_samples_per_second": 146.607,
-      "eval_gooaq_pairs_steps_per_second": 2.291,
+      "epoch": 0.9101123595505618,
+      "eval_gooaq_pairs_loss": 0.5496390461921692,
+      "eval_gooaq_pairs_runtime": 0.8735,
+      "eval_gooaq_pairs_samples_per_second": 146.532,
+      "eval_gooaq_pairs_steps_per_second": 2.29,
       "step": 162
     },
     {
-      "epoch": 0.9146083274523642,
-      "eval_paws-pos_loss": 0.052534349262714386,
-      "eval_paws-pos_runtime": 0.6835,
-      "eval_paws-pos_samples_per_second": 187.258,
+      "epoch": 0.9101123595505618,
+      "eval_paws-pos_loss": 0.024540428072214127,
+      "eval_paws-pos_runtime": 0.6836,
+      "eval_paws-pos_samples_per_second": 187.246,
       "eval_paws-pos_steps_per_second": 2.926,
       "step": 162
     },
     {
-      "epoch": 0.9315455187014821,
-      "grad_norm": 4.7817864418029785,
-      "learning_rate": 3.7500000000000003e-05,
-      "loss": 1.339,
+      "epoch": 0.9269662921348315,
+      "grad_norm": 3.270443916320801,
+      "learning_rate": 3.7525961967101216e-05,
+      "loss": 0.5395,
       "step": 165
     },
     {
-      "epoch": 0.9484827099505999,
-      "grad_norm": 4.000570774078369,
-      "learning_rate": 3.818181818181819e-05,
-      "loss": 1.1535,
+      "epoch": 0.9438202247191011,
+      "grad_norm": 2.608531951904297,
+      "learning_rate": 3.7351572532692915e-05,
+      "loss": 0.6874,
       "step": 168
     },
     {
-      "epoch": 0.9654199011997178,
-      "grad_norm": 3.5971670150756836,
-      "learning_rate": 3.8863636363636364e-05,
-      "loss": 1.1643,
+      "epoch": 0.9606741573033708,
+      "grad_norm": 2.9995415210723877,
+      "learning_rate": 3.717183680920136e-05,
+      "loss": 0.5614,
       "step": 171
     },
     {
-      "epoch": 0.9823570924488356,
-      "grad_norm": 3.6582131385803223,
-      "learning_rate": 3.954545454545455e-05,
-      "loss": 1.2221,
+      "epoch": 0.9775280898876404,
+      "grad_norm": 3.42486572265625,
+      "learning_rate": 3.698683259393809e-05,
+      "loss": 0.5812,
       "step": 174
     },
     {
-      "epoch": 0.9992942836979535,
-      "grad_norm": 4.0953898429870605,
-      "learning_rate": 3.9999477905707075e-05,
-      "loss": 1.0974,
+      "epoch": 0.9943820224719101,
+      "grad_norm": 1.8073548078536987,
+      "learning_rate": 3.679663996464331e-05,
+      "loss": 0.427,
       "step": 177
     },
     {
-      "epoch": 1.0162314749470713,
-      "grad_norm": 4.092026233673096,
-      "learning_rate": 3.999164730903481e-05,
-      "loss": 1.0984,
+      "epoch": 1.0112359550561798,
+      "grad_norm": 3.2385106086730957,
+      "learning_rate": 3.660134124482482e-05,
+      "loss": 0.4603,
       "step": 180
     },
     {
-      "epoch": 1.0331686661961892,
-      "grad_norm": 3.6480906009674072,
-      "learning_rate": 3.997442539262898e-05,
-      "loss": 1.0543,
+      "epoch": 1.0280898876404494,
+      "grad_norm": 2.8062069416046143,
+      "learning_rate": 3.640102096812488e-05,
+      "loss": 0.6493,
       "step": 183
     },
     {
-      "epoch": 1.050105857445307,
-      "grad_norm": 3.433056592941284,
-      "learning_rate": 3.99478242943326e-05,
-      "loss": 1.0994,
+      "epoch": 1.0449438202247192,
+      "grad_norm": 2.6892104148864746,
+      "learning_rate": 3.619576584173041e-05,
+      "loss": 0.6646,
       "step": 186
     },
     {
-      "epoch": 1.067043048694425,
-      "grad_norm": 3.507981777191162,
-      "learning_rate": 3.991186276234698e-05,
-      "loss": 1.0621,
+      "epoch": 1.0617977528089888,
+      "grad_norm": 2.6583099365234375,
+      "learning_rate": 3.598566470884244e-05,
+      "loss": 0.7239,
       "step": 189
     },
     {
-      "epoch": 1.067043048694425,
+      "epoch": 1.0617977528089888,
       "eval_NLI-v2_cosine_accuracy": 1.0,
       "eval_NLI-v2_dot_accuracy": 0.0,
       "eval_NLI-v2_euclidean_accuracy": 1.0,
       "eval_NLI-v2_manhattan_accuracy": 1.0,
       "eval_NLI-v2_max_accuracy": 1.0,
-      "eval_VitaminC_cosine_accuracy": 0.578125,
-      "eval_VitaminC_cosine_accuracy_threshold": 0.7840081453323364,
-      "eval_VitaminC_cosine_ap": 0.5400770399437144,
-      "eval_VitaminC_cosine_f1": 0.6577540106951871,
-      "eval_VitaminC_cosine_f1_threshold": 0.39448243379592896,
-      "eval_VitaminC_cosine_precision": 0.4900398406374502,
+      "eval_VitaminC_cosine_accuracy": 0.58984375,
+      "eval_VitaminC_cosine_accuracy_threshold": 0.8360881209373474,
+      "eval_VitaminC_cosine_ap": 0.5601848253252508,
+      "eval_VitaminC_cosine_f1": 0.6559999999999999,
+      "eval_VitaminC_cosine_f1_threshold": 0.26484909653663635,
+      "eval_VitaminC_cosine_precision": 0.4880952380952381,
       "eval_VitaminC_cosine_recall": 1.0,
-      "eval_VitaminC_dot_accuracy": 0.5625,
-      "eval_VitaminC_dot_accuracy_threshold": 323.20281982421875,
-      "eval_VitaminC_dot_ap": 0.5420016101916201,
-      "eval_VitaminC_dot_f1": 0.6575342465753424,
-      "eval_VitaminC_dot_f1_threshold": 198.04354858398438,
-      "eval_VitaminC_dot_precision": 0.49586776859504134,
-      "eval_VitaminC_dot_recall": 0.975609756097561,
-      "eval_VitaminC_euclidean_accuracy": 0.5859375,
-      "eval_VitaminC_euclidean_accuracy_threshold": 13.84214973449707,
-      "eval_VitaminC_euclidean_ap": 0.5392157650683609,
+      "eval_VitaminC_dot_accuracy": 0.58203125,
+      "eval_VitaminC_dot_accuracy_threshold": 314.279052734375,
+      "eval_VitaminC_dot_ap": 0.5513292673695236,
+      "eval_VitaminC_dot_f1": 0.6558265582655827,
+      "eval_VitaminC_dot_f1_threshold": 126.1304931640625,
+      "eval_VitaminC_dot_precision": 0.491869918699187,
+      "eval_VitaminC_dot_recall": 0.983739837398374,
+      "eval_VitaminC_euclidean_accuracy": 0.578125,
+      "eval_VitaminC_euclidean_accuracy_threshold": 15.01893424987793,
+      "eval_VitaminC_euclidean_ap": 0.5549132214851141,
       "eval_VitaminC_euclidean_f1": 0.6577540106951871,
-      "eval_VitaminC_euclidean_f1_threshold": 22.595678329467773,
+      "eval_VitaminC_euclidean_f1_threshold": 23.76571273803711,
       "eval_VitaminC_euclidean_precision": 0.4900398406374502,
       "eval_VitaminC_euclidean_recall": 1.0,
-      "eval_VitaminC_manhattan_accuracy": 0.5703125,
-      "eval_VitaminC_manhattan_accuracy_threshold": 275.1253356933594,
-      "eval_VitaminC_manhattan_ap": 0.5341380380767263,
-      "eval_VitaminC_manhattan_f1": 0.6576819407008085,
-      "eval_VitaminC_manhattan_f1_threshold": 457.04986572265625,
-      "eval_VitaminC_manhattan_precision": 0.49193548387096775,
-      "eval_VitaminC_manhattan_recall": 0.991869918699187,
-      "eval_VitaminC_max_accuracy": 0.5859375,
-      "eval_VitaminC_max_accuracy_threshold": 323.20281982421875,
-      "eval_VitaminC_max_ap": 0.5420016101916201,
+      "eval_VitaminC_manhattan_accuracy": 0.57421875,
+      "eval_VitaminC_manhattan_accuracy_threshold": 244.02972412109375,
+      "eval_VitaminC_manhattan_ap": 0.5562338006363409,
+      "eval_VitaminC_manhattan_f1": 0.6577540106951871,
+      "eval_VitaminC_manhattan_f1_threshold": 498.5762634277344,
+      "eval_VitaminC_manhattan_precision": 0.4900398406374502,
+      "eval_VitaminC_manhattan_recall": 1.0,
+      "eval_VitaminC_max_accuracy": 0.58984375,
+      "eval_VitaminC_max_accuracy_threshold": 314.279052734375,
+      "eval_VitaminC_max_ap": 0.5601848253252508,
       "eval_VitaminC_max_f1": 0.6577540106951871,
-      "eval_VitaminC_max_f1_threshold": 457.04986572265625,
-      "eval_VitaminC_max_precision": 0.49586776859504134,
+      "eval_VitaminC_max_f1_threshold": 498.5762634277344,
+      "eval_VitaminC_max_precision": 0.491869918699187,
       "eval_VitaminC_max_recall": 1.0,
-      "eval_sequential_score": 0.5420016101916201,
-      "eval_sts-test_pearson_cosine": 0.8193410747427454,
-      "eval_sts-test_pearson_dot": 0.8275444476338831,
-      "eval_sts-test_pearson_euclidean": 0.8464528142983967,
-      "eval_sts-test_pearson_manhattan": 0.8440476980962159,
-      "eval_sts-test_pearson_max": 0.8464528142983967,
-      "eval_sts-test_spearman_cosine": 0.8680272706642878,
-      "eval_sts-test_spearman_dot": 0.8555529342729671,
-      "eval_sts-test_spearman_euclidean": 0.8542457068859202,
-      "eval_sts-test_spearman_manhattan": 0.8510265117511795,
-      "eval_sts-test_spearman_max": 0.8680272706642878,
-      "eval_vitaminc-pairs_loss": 2.6755428314208984,
-      "eval_vitaminc-pairs_runtime": 1.4509,
-      "eval_vitaminc-pairs_samples_per_second": 74.437,
-      "eval_vitaminc-pairs_steps_per_second": 1.378,
+      "eval_sequential_score": 0.5601848253252508,
+      "eval_sts-test_pearson_cosine": 0.854968805652805,
+      "eval_sts-test_pearson_dot": 0.8534110565503882,
+      "eval_sts-test_pearson_euclidean": 0.8853384519331917,
+      "eval_sts-test_pearson_manhattan": 0.8864271118397893,
+      "eval_sts-test_pearson_max": 0.8864271118397893,
+      "eval_sts-test_spearman_cosine": 0.8956917253507228,
+      "eval_sts-test_spearman_dot": 0.877726389450295,
+      "eval_sts-test_spearman_euclidean": 0.8875533307992096,
+      "eval_sts-test_spearman_manhattan": 0.890112288382125,
+      "eval_sts-test_spearman_max": 0.8956917253507228,
+      "eval_vitaminc-pairs_loss": 2.3751792907714844,
+      "eval_vitaminc-pairs_runtime": 1.4524,
+      "eval_vitaminc-pairs_samples_per_second": 74.358,
+      "eval_vitaminc-pairs_steps_per_second": 1.377,
       "step": 189
     },
     {
-      "epoch": 1.067043048694425,
-      "eval_negation-triplets_loss": 1.9071491956710815,
-      "eval_negation-triplets_runtime": 0.3051,
-      "eval_negation-triplets_samples_per_second": 209.756,
-      "eval_negation-triplets_steps_per_second": 3.277,
+      "epoch": 1.0617977528089888,
+      "eval_negation-triplets_loss": 1.580323338508606,
+      "eval_negation-triplets_runtime": 0.301,
+      "eval_negation-triplets_samples_per_second": 212.649,
+      "eval_negation-triplets_steps_per_second": 3.323,
       "step": 189
     },
     {
-      "epoch": 1.067043048694425,
-      "eval_scitail-pairs-pos_loss": 0.18539850413799286,
-      "eval_scitail-pairs-pos_runtime": 0.4199,
-      "eval_scitail-pairs-pos_samples_per_second": 128.604,
-      "eval_scitail-pairs-pos_steps_per_second": 2.382,
+      "epoch": 1.0617977528089888,
+      "eval_scitail-pairs-pos_loss": 0.10438331216573715,
+      "eval_scitail-pairs-pos_runtime": 0.363,
+      "eval_scitail-pairs-pos_samples_per_second": 148.774,
+      "eval_scitail-pairs-pos_steps_per_second": 2.755,
       "step": 189
     },
     {
-      "epoch": 1.067043048694425,
-      "eval_xsum-pairs_loss": 0.38365328311920166,
-      "eval_xsum-pairs_runtime": 3.1907,
-      "eval_xsum-pairs_samples_per_second": 40.116,
-      "eval_xsum-pairs_steps_per_second": 0.627,
+      "epoch": 1.0617977528089888,
+      "eval_xsum-pairs_loss": 0.06599828600883484,
+      "eval_xsum-pairs_runtime": 3.1538,
+      "eval_xsum-pairs_samples_per_second": 40.586,
+      "eval_xsum-pairs_steps_per_second": 0.634,
       "step": 189
     },
     {
-      "epoch": 1.067043048694425,
-      "eval_sciq_pairs_loss": 0.05558515340089798,
-      "eval_sciq_pairs_runtime": 3.2891,
-      "eval_sciq_pairs_samples_per_second": 38.917,
-      "eval_sciq_pairs_steps_per_second": 0.608,
+      "epoch": 1.0617977528089888,
+      "eval_sciq_pairs_loss": 0.033071305602788925,
+      "eval_sciq_pairs_runtime": 3.2631,
+      "eval_sciq_pairs_samples_per_second": 39.227,
+      "eval_sciq_pairs_steps_per_second": 0.613,
       "step": 189
     },
     {
-      "epoch": 1.067043048694425,
-      "eval_qasc_pairs_loss": 0.40469691157341003,
-      "eval_qasc_pairs_runtime": 0.6267,
-      "eval_qasc_pairs_samples_per_second": 204.245,
-      "eval_qasc_pairs_steps_per_second": 3.191,
+      "epoch": 1.0617977528089888,
+      "eval_qasc_pairs_loss": 0.10076003521680832,
+      "eval_qasc_pairs_runtime": 0.6385,
+      "eval_qasc_pairs_samples_per_second": 200.46,
+      "eval_qasc_pairs_steps_per_second": 3.132,
       "step": 189
     },
     {
-      "epoch": 1.067043048694425,
-      "eval_openbookqa_pairs_loss": 1.0837312936782837,
-      "eval_openbookqa_pairs_runtime": 0.5765,
-      "eval_openbookqa_pairs_samples_per_second": 222.02,
-      "eval_openbookqa_pairs_steps_per_second": 3.469,
+      "epoch": 1.0617977528089888,
+      "eval_openbookqa_pairs_loss": 0.6888580918312073,
+      "eval_openbookqa_pairs_runtime": 0.6067,
+      "eval_openbookqa_pairs_samples_per_second": 210.99,
+      "eval_openbookqa_pairs_steps_per_second": 3.297,
       "step": 189
     },
     {
-      "epoch": 1.067043048694425,
-      "eval_msmarco_pairs_loss": 0.6897398233413696,
-      "eval_msmarco_pairs_runtime": 1.2918,
-      "eval_msmarco_pairs_samples_per_second": 99.089,
-      "eval_msmarco_pairs_steps_per_second": 1.548,
+      "epoch": 1.0617977528089888,
+      "eval_msmarco_pairs_loss": 0.335863322019577,
+      "eval_msmarco_pairs_runtime": 1.3206,
+      "eval_msmarco_pairs_samples_per_second": 96.922,
+      "eval_msmarco_pairs_steps_per_second": 1.514,
       "step": 189
     },
     {
-      "epoch": 1.067043048694425,
-      "eval_nq_pairs_loss": 0.9603796601295471,
-      "eval_nq_pairs_runtime": 2.3975,
-      "eval_nq_pairs_samples_per_second": 53.39,
+      "epoch": 1.0617977528089888,
+      "eval_nq_pairs_loss": 0.3690747916698456,
+      "eval_nq_pairs_runtime": 2.3983,
+      "eval_nq_pairs_samples_per_second": 53.371,
       "eval_nq_pairs_steps_per_second": 0.834,
       "step": 189
     },
     {
-      "epoch": 1.067043048694425,
-      "eval_trivia_pairs_loss": 1.200446605682373,
-      "eval_trivia_pairs_runtime": 4.4582,
-      "eval_trivia_pairs_samples_per_second": 28.711,
-      "eval_trivia_pairs_steps_per_second": 0.449,
+      "epoch": 1.0617977528089888,
+      "eval_trivia_pairs_loss": 0.881881594657898,
+      "eval_trivia_pairs_runtime": 4.4261,
+      "eval_trivia_pairs_samples_per_second": 28.919,
+      "eval_trivia_pairs_steps_per_second": 0.452,
       "step": 189
     },
     {
-      "epoch": 1.067043048694425,
-      "eval_gooaq_pairs_loss": 1.0353316068649292,
-      "eval_gooaq_pairs_runtime": 0.8765,
-      "eval_gooaq_pairs_samples_per_second": 146.042,
-      "eval_gooaq_pairs_steps_per_second": 2.282,
+      "epoch": 1.0617977528089888,
+      "eval_gooaq_pairs_loss": 0.5453925728797913,
+      "eval_gooaq_pairs_runtime": 0.8775,
+      "eval_gooaq_pairs_samples_per_second": 145.871,
+      "eval_gooaq_pairs_steps_per_second": 2.279,
       "step": 189
     },
     {
-      "epoch": 1.067043048694425,
-      "eval_paws-pos_loss": 0.042069558054208755,
-      "eval_paws-pos_runtime": 0.6909,
-      "eval_paws-pos_samples_per_second": 185.263,
-      "eval_paws-pos_steps_per_second": 2.895,
+      "epoch": 1.0617977528089888,
+      "eval_paws-pos_loss": 0.024841103702783585,
+      "eval_paws-pos_runtime": 0.6851,
+      "eval_paws-pos_samples_per_second": 186.83,
+      "eval_paws-pos_steps_per_second": 2.919,
       "step": 189
     },
     {
-      "epoch": 1.0839802399435428,
-      "grad_norm": 2.979419469833374,
-      "learning_rate": 3.986656614201813e-05,
-      "loss": 0.8724,
+      "epoch": 1.0786516853932584,
+      "grad_norm": 2.942894220352173,
+      "learning_rate": 3.5770808510220957e-05,
+      "loss": 0.7593,
       "step": 192
     },
     {
-      "epoch": 1.1009174311926606,
-      "grad_norm": 2.835219144821167,
-      "learning_rate": 3.981196635797361e-05,
-      "loss": 0.9381,
+      "epoch": 1.095505617977528,
+      "grad_norm": 2.636993169784546,
+      "learning_rate": 3.5551290244821855e-05,
+      "loss": 0.6877,
       "step": 195
     },
     {
-      "epoch": 1.1178546224417785,
-      "grad_norm": 3.6650869846343994,
-      "learning_rate": 3.974810189162238e-05,
-      "loss": 0.9617,
+      "epoch": 1.1123595505617978,
+      "grad_norm": 2.7987968921661377,
+      "learning_rate": 3.5327204929543e-05,
+      "loss": 0.5482,
       "step": 198
     },
     {
-      "epoch": 1.1347918136908963,
-      "grad_norm": 4.188896656036377,
-      "learning_rate": 3.967501775403343e-05,
-      "loss": 1.0139,
+      "epoch": 1.1292134831460674,
+      "grad_norm": 2.3775548934936523,
+      "learning_rate": 3.509864955809687e-05,
+      "loss": 0.6047,
       "step": 201
     },
     {
-      "epoch": 1.1517290049400142,
-      "grad_norm": 3.1624915599823,
-      "learning_rate": 3.959276545421244e-05,
-      "loss": 1.1073,
+      "epoch": 1.146067415730337,
+      "grad_norm": 1.982017993927002,
+      "learning_rate": 3.4865723059027493e-05,
+      "loss": 0.4358,
       "step": 204
     },
     {
-      "epoch": 1.168666196189132,
-      "grad_norm": 3.245002508163452,
-      "learning_rate": 3.950140296279871e-05,
-      "loss": 0.8365,
+      "epoch": 1.1629213483146068,
+      "grad_norm": 1.5380574464797974,
+      "learning_rate": 3.462852625288999e-05,
+      "loss": 0.3343,
       "step": 207
     },
     {
-      "epoch": 1.18560338743825,
-      "grad_norm": 4.376185894012451,
-      "learning_rate": 3.9400994671208e-05,
-      "loss": 1.1012,
+      "epoch": 1.1797752808988764,
+      "grad_norm": 3.152486562728882,
+      "learning_rate": 3.438716180861106e-05,
+      "loss": 0.5624,
       "step": 210
     },
     {
-      "epoch": 1.2025405786873677,
-      "grad_norm": 3.236583948135376,
-      "learning_rate": 3.9291611346250066e-05,
-      "loss": 1.0016,
+      "epoch": 1.196629213483146,
+      "grad_norm": 2.410505533218384,
+      "learning_rate": 3.414173419904956e-05,
+      "loss": 0.4578,
       "step": 213
     }
   ],
   "logging_steps": 3,
-  "max_steps": 531,
+  "max_steps": 534,
   "num_input_tokens_seen": 0,
   "num_train_epochs": 3,
   "save_steps": 107,
@@ -1553,7 +1553,7 @@
     }
   },
   "total_flos": 0.0,
-  "train_batch_size": 160,
+  "train_batch_size": 320,
   "trial_name": null,
   "trial_params": null
 }