Training in progress, epoch 1

model.safetensors

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:fb19d8f5f60ec559ca8b0f113d6118526dbcc7804bd492ae19376c21a4b3b51e
+oid sha256:d9a9413e0558b2b03f34ffaeef8d03315ff63af1209bdfcd55acf4ce04908097
 size 438100144

run-1/checkpoint-11995/config.json

@@ -0,0 +1,125 @@
+{
+  "architectures": [
+    "BertForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "classifier_dropout": null,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "LABEL_0",
+    "1": "LABEL_1",
+    "2": "LABEL_2",
+    "3": "LABEL_3",
+    "4": "LABEL_4",
+    "5": "LABEL_5",
+    "6": "LABEL_6",
+    "7": "LABEL_7",
+    "8": "LABEL_8",
+    "9": "LABEL_9",
+    "10": "LABEL_10",
+    "11": "LABEL_11",
+    "12": "LABEL_12",
+    "13": "LABEL_13",
+    "14": "LABEL_14",
+    "15": "LABEL_15",
+    "16": "LABEL_16",
+    "17": "LABEL_17",
+    "18": "LABEL_18",
+    "19": "LABEL_19",
+    "20": "LABEL_20",
+    "21": "LABEL_21",
+    "22": "LABEL_22",
+    "23": "LABEL_23",
+    "24": "LABEL_24",
+    "25": "LABEL_25",
+    "26": "LABEL_26",
+    "27": "LABEL_27",
+    "28": "LABEL_28",
+    "29": "LABEL_29",
+    "30": "LABEL_30",
+    "31": "LABEL_31",
+    "32": "LABEL_32",
+    "33": "LABEL_33",
+    "34": "LABEL_34",
+    "35": "LABEL_35",
+    "36": "LABEL_36",
+    "37": "LABEL_37",
+    "38": "LABEL_38",
+    "39": "LABEL_39",
+    "40": "LABEL_40",
+    "41": "LABEL_41",
+    "42": "LABEL_42",
+    "43": "LABEL_43",
+    "44": "LABEL_44",
+    "45": "LABEL_45",
+    "46": "LABEL_46",
+    "47": "LABEL_47"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "LABEL_0": 0,
+    "LABEL_1": 1,
+    "LABEL_10": 10,
+    "LABEL_11": 11,
+    "LABEL_12": 12,
+    "LABEL_13": 13,
+    "LABEL_14": 14,
+    "LABEL_15": 15,
+    "LABEL_16": 16,
+    "LABEL_17": 17,
+    "LABEL_18": 18,
+    "LABEL_19": 19,
+    "LABEL_2": 2,
+    "LABEL_20": 20,
+    "LABEL_21": 21,
+    "LABEL_22": 22,
+    "LABEL_23": 23,
+    "LABEL_24": 24,
+    "LABEL_25": 25,
+    "LABEL_26": 26,
+    "LABEL_27": 27,
+    "LABEL_28": 28,
+    "LABEL_29": 29,
+    "LABEL_3": 3,
+    "LABEL_30": 30,
+    "LABEL_31": 31,
+    "LABEL_32": 32,
+    "LABEL_33": 33,
+    "LABEL_34": 34,
+    "LABEL_35": 35,
+    "LABEL_36": 36,
+    "LABEL_37": 37,
+    "LABEL_38": 38,
+    "LABEL_39": 39,
+    "LABEL_4": 4,
+    "LABEL_40": 40,
+    "LABEL_41": 41,
+    "LABEL_42": 42,
+    "LABEL_43": 43,
+    "LABEL_44": 44,
+    "LABEL_45": 45,
+    "LABEL_46": 46,
+    "LABEL_47": 47,
+    "LABEL_5": 5,
+    "LABEL_6": 6,
+    "LABEL_7": 7,
+    "LABEL_8": 8,
+    "LABEL_9": 9
+  },
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "torch_dtype": "float32",
+  "transformers_version": "4.50.3",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 30522
+}

run-1/checkpoint-11995/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a986b998b035933a22ebb80fa1d2983a26d968904eac727ae2c6659a283ad9d8
+size 438100144

run-1/checkpoint-11995/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7a153653a8f9db4121420ac8d5bd31548e3e34d6496d889fe67d95642d32724a
+size 876321402

run-1/checkpoint-11995/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8b3f51a81800533b5f405ae674a60a515fe53b8ffe03db1840c5ab046ef1725d
+size 14244

run-1/checkpoint-11995/scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:2eadb731d755c8a0c25e025041a12f77a9bb9d88de47337097d3d598fe20f0e4
+size 988

run-1/checkpoint-11995/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:6b16952129b4bb2cd68b5609aeec289e775814b8205dc2fc50f03261fbda4de0
+size 1064

run-1/checkpoint-11995/trainer_state.json

@@ -0,0 +1,1050 @@
+{
+  "best_global_step": 11995,
+  "best_metric": 0.9374392372783462,
+  "best_model_checkpoint": "./results/run-1/checkpoint-11995",
+  "epoch": 5.0,
+  "eval_steps": 500,
+  "global_step": 11995,
+  "is_hyper_param_search": true,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.041684035014589414,
+      "grad_norm": 2.362386465072632,
+      "learning_rate": 9.900000000000002e-06,
+      "loss": 1.3944,
+      "step": 100
+    },
+    {
+      "epoch": 0.08336807002917883,
+      "grad_norm": 0.7483571767807007,
+      "learning_rate": 1.9900000000000003e-05,
+      "loss": 0.5576,
+      "step": 200
+    },
+    {
+      "epoch": 0.12505210504376824,
+      "grad_norm": 0.8563070297241211,
+      "learning_rate": 2.9900000000000002e-05,
+      "loss": 0.4502,
+      "step": 300
+    },
+    {
+      "epoch": 0.16673614005835766,
+      "grad_norm": 0.7200814485549927,
+      "learning_rate": 3.99e-05,
+      "loss": 0.388,
+      "step": 400
+    },
+    {
+      "epoch": 0.20842017507294705,
+      "grad_norm": 1.4553083181381226,
+      "learning_rate": 4.99e-05,
+      "loss": 0.3508,
+      "step": 500
+    },
+    {
+      "epoch": 0.25010421008753647,
+      "grad_norm": 0.9915399551391602,
+      "learning_rate": 4.978927203065135e-05,
+      "loss": 0.3186,
+      "step": 600
+    },
+    {
+      "epoch": 0.29178824510212586,
+      "grad_norm": 0.9978141188621521,
+      "learning_rate": 4.957641549595573e-05,
+      "loss": 0.3087,
+      "step": 700
+    },
+    {
+      "epoch": 0.3334722801167153,
+      "grad_norm": 1.3709461688995361,
+      "learning_rate": 4.936355896126011e-05,
+      "loss": 0.2976,
+      "step": 800
+    },
+    {
+      "epoch": 0.3751563151313047,
+      "grad_norm": 1.82745361328125,
+      "learning_rate": 4.91507024265645e-05,
+      "loss": 0.28,
+      "step": 900
+    },
+    {
+      "epoch": 0.4168403501458941,
+      "grad_norm": 1.2072925567626953,
+      "learning_rate": 4.893784589186888e-05,
+      "loss": 0.2751,
+      "step": 1000
+    },
+    {
+      "epoch": 0.45852438516048355,
+      "grad_norm": 1.0433937311172485,
+      "learning_rate": 4.872498935717327e-05,
+      "loss": 0.2712,
+      "step": 1100
+    },
+    {
+      "epoch": 0.5002084201750729,
+      "grad_norm": 1.1097549200057983,
+      "learning_rate": 4.851213282247765e-05,
+      "loss": 0.2909,
+      "step": 1200
+    },
+    {
+      "epoch": 0.5418924551896623,
+      "grad_norm": 1.2917771339416504,
+      "learning_rate": 4.8299276287782034e-05,
+      "loss": 0.2501,
+      "step": 1300
+    },
+    {
+      "epoch": 0.5835764902042517,
+      "grad_norm": 0.9812286496162415,
+      "learning_rate": 4.808641975308642e-05,
+      "loss": 0.2621,
+      "step": 1400
+    },
+    {
+      "epoch": 0.6252605252188412,
+      "grad_norm": 1.1786612272262573,
+      "learning_rate": 4.7873563218390804e-05,
+      "loss": 0.2435,
+      "step": 1500
+    },
+    {
+      "epoch": 0.6669445602334306,
+      "grad_norm": 1.4759145975112915,
+      "learning_rate": 4.766070668369519e-05,
+      "loss": 0.2398,
+      "step": 1600
+    },
+    {
+      "epoch": 0.70862859524802,
+      "grad_norm": 1.5548242330551147,
+      "learning_rate": 4.744785014899958e-05,
+      "loss": 0.2432,
+      "step": 1700
+    },
+    {
+      "epoch": 0.7503126302626094,
+      "grad_norm": 1.9005539417266846,
+      "learning_rate": 4.723499361430396e-05,
+      "loss": 0.2351,
+      "step": 1800
+    },
+    {
+      "epoch": 0.7919966652771988,
+      "grad_norm": 1.5831588506698608,
+      "learning_rate": 4.702213707960835e-05,
+      "loss": 0.2378,
+      "step": 1900
+    },
+    {
+      "epoch": 0.8336807002917882,
+      "grad_norm": 1.5482088327407837,
+      "learning_rate": 4.680928054491273e-05,
+      "loss": 0.2271,
+      "step": 2000
+    },
+    {
+      "epoch": 0.8753647353063777,
+      "grad_norm": 1.1437913179397583,
+      "learning_rate": 4.6596424010217114e-05,
+      "loss": 0.231,
+      "step": 2100
+    },
+    {
+      "epoch": 0.9170487703209671,
+      "grad_norm": 0.7985406517982483,
+      "learning_rate": 4.63835674755215e-05,
+      "loss": 0.228,
+      "step": 2200
+    },
+    {
+      "epoch": 0.9587328053355565,
+      "grad_norm": 1.9323142766952515,
+      "learning_rate": 4.6170710940825884e-05,
+      "loss": 0.2198,
+      "step": 2300
+    },
+    {
+      "epoch": 1.0,
+      "eval_administration_accuracy": 0.9322493224932249,
+      "eval_administration_f1": 0.9122911143651031,
+      "eval_corruption_accuracy": 0.9560141755263707,
+      "eval_corruption_f1": 0.9517304794081561,
+      "eval_democracy_accuracy": 0.9505941213258287,
+      "eval_democracy_f1": 0.9401838742047087,
+      "eval_development_accuracy": 0.9118198874296435,
+      "eval_development_f1": 0.9003553411809495,
+      "eval_economy_accuracy": 0.9264123410464874,
+      "eval_economy_f1": 0.9202465382357672,
+      "eval_education_accuracy": 0.9643527204502814,
+      "eval_education_f1": 0.9615509617901187,
+      "eval_environment_accuracy": 0.9799874921826142,
+      "eval_environment_f1": 0.9786142985651582,
+      "eval_instability_accuracy": 0.9417344173441734,
+      "eval_instability_f1": 0.9340921018628459,
+      "eval_leadership_accuracy": 0.8085261621846987,
+      "eval_leadership_f1": 0.7858417402549625,
+      "eval_loss": 0.21971678733825684,
+      "eval_overall_accuracy": 0.9324317281634354,
+      "eval_overall_f1": 0.9244250028140656,
+      "eval_race_accuracy": 0.9508025849489264,
+      "eval_race_f1": 0.9481714649378918,
+      "eval_religion_accuracy": 0.9530956848030019,
+      "eval_religion_f1": 0.9518106778632144,
+      "eval_runtime": 10.9084,
+      "eval_safety_accuracy": 0.9135918282259746,
+      "eval_safety_f1": 0.9082114410999095,
+      "eval_samples_per_second": 879.505,
+      "eval_steps_per_second": 55.003,
+      "step": 2399
+    },
+    {
+      "epoch": 1.0004168403501459,
+      "grad_norm": 1.2817295789718628,
+      "learning_rate": 4.5957854406130265e-05,
+      "loss": 0.2397,
+      "step": 2400
+    },
+    {
+      "epoch": 1.0421008753647354,
+      "grad_norm": 1.3623204231262207,
+      "learning_rate": 4.5744997871434654e-05,
+      "loss": 0.1885,
+      "step": 2500
+    },
+    {
+      "epoch": 1.0837849103793247,
+      "grad_norm": 1.2534806728363037,
+      "learning_rate": 4.553214133673904e-05,
+      "loss": 0.1843,
+      "step": 2600
+    },
+    {
+      "epoch": 1.1254689453939142,
+      "grad_norm": 1.2860488891601562,
+      "learning_rate": 4.5319284802043424e-05,
+      "loss": 0.1817,
+      "step": 2700
+    },
+    {
+      "epoch": 1.1671529804085035,
+      "grad_norm": 1.2887414693832397,
+      "learning_rate": 4.510642826734781e-05,
+      "loss": 0.1798,
+      "step": 2800
+    },
+    {
+      "epoch": 1.208837015423093,
+      "grad_norm": 1.973421335220337,
+      "learning_rate": 4.4893571732652194e-05,
+      "loss": 0.1878,
+      "step": 2900
+    },
+    {
+      "epoch": 1.2505210504376825,
+      "grad_norm": 0.7910940051078796,
+      "learning_rate": 4.468071519795658e-05,
+      "loss": 0.182,
+      "step": 3000
+    },
+    {
+      "epoch": 1.2922050854522718,
+      "grad_norm": 1.7422680854797363,
+      "learning_rate": 4.4467858663260964e-05,
+      "loss": 0.1833,
+      "step": 3100
+    },
+    {
+      "epoch": 1.3338891204668613,
+      "grad_norm": 1.1493451595306396,
+      "learning_rate": 4.4255002128565345e-05,
+      "loss": 0.1766,
+      "step": 3200
+    },
+    {
+      "epoch": 1.3755731554814505,
+      "grad_norm": 2.179731845855713,
+      "learning_rate": 4.4042145593869734e-05,
+      "loss": 0.1785,
+      "step": 3300
+    },
+    {
+      "epoch": 1.41725719049604,
+      "grad_norm": 1.5773117542266846,
+      "learning_rate": 4.3829289059174115e-05,
+      "loss": 0.1901,
+      "step": 3400
+    },
+    {
+      "epoch": 1.4589412255106295,
+      "grad_norm": 2.4681854248046875,
+      "learning_rate": 4.3616432524478504e-05,
+      "loss": 0.1833,
+      "step": 3500
+    },
+    {
+      "epoch": 1.5006252605252188,
+      "grad_norm": 2.391803503036499,
+      "learning_rate": 4.340357598978289e-05,
+      "loss": 0.1677,
+      "step": 3600
+    },
+    {
+      "epoch": 1.5423092955398081,
+      "grad_norm": 0.8182645440101624,
+      "learning_rate": 4.3190719455087274e-05,
+      "loss": 0.1666,
+      "step": 3700
+    },
+    {
+      "epoch": 1.5839933305543976,
+      "grad_norm": 1.2136952877044678,
+      "learning_rate": 4.297786292039166e-05,
+      "loss": 0.1787,
+      "step": 3800
+    },
+    {
+      "epoch": 1.6256773655689871,
+      "grad_norm": 2.435816764831543,
+      "learning_rate": 4.2765006385696044e-05,
+      "loss": 0.1678,
+      "step": 3900
+    },
+    {
+      "epoch": 1.6673614005835766,
+      "grad_norm": 1.5172932147979736,
+      "learning_rate": 4.2552149851000425e-05,
+      "loss": 0.1779,
+      "step": 4000
+    },
+    {
+      "epoch": 1.709045435598166,
+      "grad_norm": 1.4677119255065918,
+      "learning_rate": 4.2339293316304814e-05,
+      "loss": 0.1824,
+      "step": 4100
+    },
+    {
+      "epoch": 1.7507294706127552,
+      "grad_norm": 1.8327163457870483,
+      "learning_rate": 4.2126436781609195e-05,
+      "loss": 0.1719,
+      "step": 4200
+    },
+    {
+      "epoch": 1.7924135056273447,
+      "grad_norm": 1.2324726581573486,
+      "learning_rate": 4.1913580246913584e-05,
+      "loss": 0.1592,
+      "step": 4300
+    },
+    {
+      "epoch": 1.8340975406419342,
+      "grad_norm": 3.6276469230651855,
+      "learning_rate": 4.1700723712217965e-05,
+      "loss": 0.1636,
+      "step": 4400
+    },
+    {
+      "epoch": 1.8757815756565237,
+      "grad_norm": 0.940485417842865,
+      "learning_rate": 4.148786717752235e-05,
+      "loss": 0.1639,
+      "step": 4500
+    },
+    {
+      "epoch": 1.917465610671113,
+      "grad_norm": 2.002577781677246,
+      "learning_rate": 4.127501064282674e-05,
+      "loss": 0.1756,
+      "step": 4600
+    },
+    {
+      "epoch": 1.9591496456857023,
+      "grad_norm": 2.6062567234039307,
+      "learning_rate": 4.1062154108131124e-05,
+      "loss": 0.1729,
+      "step": 4700
+    },
+    {
+      "epoch": 2.0,
+      "eval_administration_accuracy": 0.9374609130706691,
+      "eval_administration_f1": 0.9302102302828306,
+      "eval_corruption_accuracy": 0.9603919116114238,
+      "eval_corruption_f1": 0.9585025895331236,
+      "eval_democracy_accuracy": 0.9562226391494684,
+      "eval_democracy_f1": 0.9505005443286667,
+      "eval_development_accuracy": 0.9199499687304565,
+      "eval_development_f1": 0.9145013995891959,
+      "eval_economy_accuracy": 0.9319366270585783,
+      "eval_economy_f1": 0.9302256375627648,
+      "eval_education_accuracy": 0.9651865749426725,
+      "eval_education_f1": 0.9641689686262792,
+      "eval_environment_accuracy": 0.9806128830519074,
+      "eval_environment_f1": 0.9798389855050841,
+      "eval_instability_accuracy": 0.9244319366270586,
+      "eval_instability_f1": 0.9301250908302109,
+      "eval_leadership_accuracy": 0.8395872420262664,
+      "eval_leadership_f1": 0.830858580138579,
+      "eval_loss": 0.20445191860198975,
+      "eval_overall_accuracy": 0.9366791744840525,
+      "eval_overall_f1": 0.9342598776015526,
+      "eval_race_accuracy": 0.9545549301646863,
+      "eval_race_f1": 0.9531323667533257,
+      "eval_religion_accuracy": 0.9575776526996039,
+      "eval_religion_f1": 0.9560680494194458,
+      "eval_runtime": 10.9745,
+      "eval_safety_accuracy": 0.9122368146758391,
+      "eval_safety_f1": 0.9129860886491253,
+      "eval_samples_per_second": 874.208,
+      "eval_steps_per_second": 54.672,
+      "step": 4798
+    },
+    {
+      "epoch": 2.0008336807002918,
+      "grad_norm": 1.0561970472335815,
+      "learning_rate": 4.085142613878246e-05,
+      "loss": 0.174,
+      "step": 4800
+    },
+    {
+      "epoch": 2.0425177157148813,
+      "grad_norm": 0.9414407014846802,
+      "learning_rate": 4.0638569604086845e-05,
+      "loss": 0.1259,
+      "step": 4900
+    },
+    {
+      "epoch": 2.084201750729471,
+      "grad_norm": 1.0318489074707031,
+      "learning_rate": 4.042571306939123e-05,
+      "loss": 0.1258,
+      "step": 5000
+    },
+    {
+      "epoch": 2.12588578574406,
+      "grad_norm": 2.4367470741271973,
+      "learning_rate": 4.0212856534695615e-05,
+      "loss": 0.1186,
+      "step": 5100
+    },
+    {
+      "epoch": 2.1675698207586493,
+      "grad_norm": 1.8129918575286865,
+      "learning_rate": 4e-05,
+      "loss": 0.1243,
+      "step": 5200
+    },
+    {
+      "epoch": 2.209253855773239,
+      "grad_norm": 2.235281467437744,
+      "learning_rate": 3.978714346530439e-05,
+      "loss": 0.1169,
+      "step": 5300
+    },
+    {
+      "epoch": 2.2509378907878284,
+      "grad_norm": 3.7830569744110107,
+      "learning_rate": 3.957428693060877e-05,
+      "loss": 0.1177,
+      "step": 5400
+    },
+    {
+      "epoch": 2.292621925802418,
+      "grad_norm": 1.2535426616668701,
+      "learning_rate": 3.9361430395913155e-05,
+      "loss": 0.1204,
+      "step": 5500
+    },
+    {
+      "epoch": 2.334305960817007,
+      "grad_norm": 1.443596601486206,
+      "learning_rate": 3.914857386121754e-05,
+      "loss": 0.1187,
+      "step": 5600
+    },
+    {
+      "epoch": 2.3759899958315964,
+      "grad_norm": 1.0241508483886719,
+      "learning_rate": 3.8935717326521925e-05,
+      "loss": 0.1222,
+      "step": 5700
+    },
+    {
+      "epoch": 2.417674030846186,
+      "grad_norm": 1.923004150390625,
+      "learning_rate": 3.8722860791826307e-05,
+      "loss": 0.1246,
+      "step": 5800
+    },
+    {
+      "epoch": 2.4593580658607754,
+      "grad_norm": 0.7686344385147095,
+      "learning_rate": 3.8510004257130695e-05,
+      "loss": 0.13,
+      "step": 5900
+    },
+    {
+      "epoch": 2.501042100875365,
+      "grad_norm": 1.762736439704895,
+      "learning_rate": 3.8297147722435077e-05,
+      "loss": 0.1263,
+      "step": 6000
+    },
+    {
+      "epoch": 2.542726135889954,
+      "grad_norm": 1.9006534814834595,
+      "learning_rate": 3.8084291187739465e-05,
+      "loss": 0.1152,
+      "step": 6100
+    },
+    {
+      "epoch": 2.5844101709045435,
+      "grad_norm": 2.003476858139038,
+      "learning_rate": 3.787143465304385e-05,
+      "loss": 0.1318,
+      "step": 6200
+    },
+    {
+      "epoch": 2.626094205919133,
+      "grad_norm": 1.8864182233810425,
+      "learning_rate": 3.7658578118348235e-05,
+      "loss": 0.1296,
+      "step": 6300
+    },
+    {
+      "epoch": 2.6677782409337225,
+      "grad_norm": 1.567968726158142,
+      "learning_rate": 3.744572158365262e-05,
+      "loss": 0.1292,
+      "step": 6400
+    },
+    {
+      "epoch": 2.7094622759483116,
+      "grad_norm": 1.7688707113265991,
+      "learning_rate": 3.7232865048957005e-05,
+      "loss": 0.1258,
+      "step": 6500
+    },
+    {
+      "epoch": 2.751146310962901,
+      "grad_norm": 2.138125419616699,
+      "learning_rate": 3.7020008514261387e-05,
+      "loss": 0.1208,
+      "step": 6600
+    },
+    {
+      "epoch": 2.7928303459774906,
+      "grad_norm": 3.3148305416107178,
+      "learning_rate": 3.6807151979565775e-05,
+      "loss": 0.1253,
+      "step": 6700
+    },
+    {
+      "epoch": 2.83451438099208,
+      "grad_norm": 1.8787152767181396,
+      "learning_rate": 3.6594295444870157e-05,
+      "loss": 0.1161,
+      "step": 6800
+    },
+    {
+      "epoch": 2.8761984160066696,
+      "grad_norm": 0.7995413541793823,
+      "learning_rate": 3.6381438910174545e-05,
+      "loss": 0.1196,
+      "step": 6900
+    },
+    {
+      "epoch": 2.917882451021259,
+      "grad_norm": 1.973639965057373,
+      "learning_rate": 3.6168582375478927e-05,
+      "loss": 0.1227,
+      "step": 7000
+    },
+    {
+      "epoch": 2.959566486035848,
+      "grad_norm": 1.0217853784561157,
+      "learning_rate": 3.5955725840783315e-05,
+      "loss": 0.125,
+      "step": 7100
+    },
+    {
+      "epoch": 3.0,
+      "eval_administration_accuracy": 0.9373566812591203,
+      "eval_administration_f1": 0.9333482432043958,
+      "eval_corruption_accuracy": 0.9609130706691682,
+      "eval_corruption_f1": 0.958848962871965,
+      "eval_democracy_accuracy": 0.9567437982072129,
+      "eval_democracy_f1": 0.9532298494465137,
+      "eval_development_accuracy": 0.9128622055451324,
+      "eval_development_f1": 0.9113351170841872,
+      "eval_economy_accuracy": 0.9317281634354805,
+      "eval_economy_f1": 0.9292396676492435,
+      "eval_education_accuracy": 0.965603502188868,
+      "eval_education_f1": 0.9637981071832747,
+      "eval_environment_accuracy": 0.9804044194288096,
+      "eval_environment_f1": 0.9788214035384287,
+      "eval_instability_accuracy": 0.9490306441525954,
+      "eval_instability_f1": 0.9469525951807101,
+      "eval_leadership_accuracy": 0.8446946007921617,
+      "eval_leadership_f1": 0.840338137042441,
+      "eval_loss": 0.20486579835414886,
+      "eval_overall_accuracy": 0.9398235007991106,
+      "eval_overall_f1": 0.9372835834491008,
+      "eval_race_accuracy": 0.9540337711069419,
+      "eval_race_f1": 0.9520143591935892,
+      "eval_religion_accuracy": 0.959349593495935,
+      "eval_religion_f1": 0.9575545048864775,
+      "eval_runtime": 10.9937,
+      "eval_safety_accuracy": 0.9251615593079008,
+      "eval_safety_f1": 0.9219220541079818,
+      "eval_samples_per_second": 872.682,
+      "eval_steps_per_second": 54.577,
+      "step": 7197
+    },
+    {
+      "epoch": 3.0012505210504377,
+      "grad_norm": 0.8939034938812256,
+      "learning_rate": 3.5744997871434655e-05,
+      "loss": 0.1224,
+      "step": 7200
+    },
+    {
+      "epoch": 3.042934556065027,
+      "grad_norm": 2.387557029724121,
+      "learning_rate": 3.5532141336739036e-05,
+      "loss": 0.0924,
+      "step": 7300
+    },
+    {
+      "epoch": 3.0846185910796167,
+      "grad_norm": 0.7587341666221619,
+      "learning_rate": 3.5319284802043425e-05,
+      "loss": 0.0838,
+      "step": 7400
+    },
+    {
+      "epoch": 3.1263026260942057,
+      "grad_norm": 2.06026291847229,
+      "learning_rate": 3.5106428267347806e-05,
+      "loss": 0.0894,
+      "step": 7500
+    },
+    {
+      "epoch": 3.1679866611087952,
+      "grad_norm": 2.001525402069092,
+      "learning_rate": 3.4893571732652195e-05,
+      "loss": 0.0913,
+      "step": 7600
+    },
+    {
+      "epoch": 3.2096706961233847,
+      "grad_norm": 2.0007410049438477,
+      "learning_rate": 3.468071519795658e-05,
+      "loss": 0.0891,
+      "step": 7700
+    },
+    {
+      "epoch": 3.2513547311379742,
+      "grad_norm": 1.6416945457458496,
+      "learning_rate": 3.4467858663260965e-05,
+      "loss": 0.0919,
+      "step": 7800
+    },
+    {
+      "epoch": 3.2930387661525637,
+      "grad_norm": 1.653064489364624,
+      "learning_rate": 3.425500212856535e-05,
+      "loss": 0.0932,
+      "step": 7900
+    },
+    {
+      "epoch": 3.334722801167153,
+      "grad_norm": 1.7365379333496094,
+      "learning_rate": 3.4042145593869735e-05,
+      "loss": 0.0898,
+      "step": 8000
+    },
+    {
+      "epoch": 3.3764068361817423,
+      "grad_norm": 1.561340093612671,
+      "learning_rate": 3.3829289059174116e-05,
+      "loss": 0.0866,
+      "step": 8100
+    },
+    {
+      "epoch": 3.418090871196332,
+      "grad_norm": 0.7200934886932373,
+      "learning_rate": 3.3616432524478505e-05,
+      "loss": 0.0844,
+      "step": 8200
+    },
+    {
+      "epoch": 3.4597749062109213,
+      "grad_norm": 0.8205280303955078,
+      "learning_rate": 3.3405704555129844e-05,
+      "loss": 0.0914,
+      "step": 8300
+    },
+    {
+      "epoch": 3.5014589412255104,
+      "grad_norm": 1.7642834186553955,
+      "learning_rate": 3.319284802043423e-05,
+      "loss": 0.0873,
+      "step": 8400
+    },
+    {
+      "epoch": 3.5431429762401,
+      "grad_norm": 2.3324527740478516,
+      "learning_rate": 3.2979991485738614e-05,
+      "loss": 0.0808,
+      "step": 8500
+    },
+    {
+      "epoch": 3.5848270112546894,
+      "grad_norm": 1.7013909816741943,
+      "learning_rate": 3.2767134951043e-05,
+      "loss": 0.094,
+      "step": 8600
+    },
+    {
+      "epoch": 3.626511046269279,
+      "grad_norm": 4.724839210510254,
+      "learning_rate": 3.2554278416347384e-05,
+      "loss": 0.0878,
+      "step": 8700
+    },
+    {
+      "epoch": 3.6681950812838684,
+      "grad_norm": 3.1913349628448486,
+      "learning_rate": 3.2341421881651766e-05,
+      "loss": 0.0918,
+      "step": 8800
+    },
+    {
+      "epoch": 3.709879116298458,
+      "grad_norm": 2.2082605361938477,
+      "learning_rate": 3.2128565346956154e-05,
+      "loss": 0.0846,
+      "step": 8900
+    },
+    {
+      "epoch": 3.751563151313047,
+      "grad_norm": 2.2428789138793945,
+      "learning_rate": 3.1915708812260536e-05,
+      "loss": 0.0878,
+      "step": 9000
+    },
+    {
+      "epoch": 3.7932471863276365,
+      "grad_norm": 2.0431671142578125,
+      "learning_rate": 3.170285227756492e-05,
+      "loss": 0.0964,
+      "step": 9100
+    },
+    {
+      "epoch": 3.834931221342226,
+      "grad_norm": 1.9629932641983032,
+      "learning_rate": 3.1489995742869306e-05,
+      "loss": 0.0969,
+      "step": 9200
+    },
+    {
+      "epoch": 3.8766152563568155,
+      "grad_norm": 1.0876961946487427,
+      "learning_rate": 3.1277139208173694e-05,
+      "loss": 0.0864,
+      "step": 9300
+    },
+    {
+      "epoch": 3.9182992913714045,
+      "grad_norm": 2.003713607788086,
+      "learning_rate": 3.1064282673478076e-05,
+      "loss": 0.0843,
+      "step": 9400
+    },
+    {
+      "epoch": 3.959983326385994,
+      "grad_norm": 1.3784735202789307,
+      "learning_rate": 3.0851426138782464e-05,
+      "loss": 0.0866,
+      "step": 9500
+    },
+    {
+      "epoch": 4.0,
+      "eval_administration_accuracy": 0.9397540129247446,
+      "eval_administration_f1": 0.9326250155599194,
+      "eval_corruption_accuracy": 0.9600792161767772,
+      "eval_corruption_f1": 0.9585031567143241,
+      "eval_democracy_accuracy": 0.9542422347300397,
+      "eval_democracy_f1": 0.9522572809405702,
+      "eval_development_accuracy": 0.9216176777152387,
+      "eval_development_f1": 0.9176939504856475,
+      "eval_economy_accuracy": 0.9336043360433605,
+      "eval_economy_f1": 0.9317409802418596,
+      "eval_education_accuracy": 0.968313529289139,
+      "eval_education_f1": 0.9659804103888047,
+      "eval_environment_accuracy": 0.9789451740671253,
+      "eval_environment_f1": 0.9783634505314646,
+      "eval_instability_accuracy": 0.9435063581405044,
+      "eval_instability_f1": 0.9430673845492272,
+      "eval_leadership_accuracy": 0.8415676464456953,
+      "eval_leadership_f1": 0.8394569407264164,
+      "eval_loss": 0.22299017012119293,
+      "eval_overall_accuracy": 0.9397540129247446,
+      "eval_overall_f1": 0.9374286167544477,
+      "eval_race_accuracy": 0.9551803210339795,
+      "eval_race_f1": 0.9543972425939777,
+      "eval_religion_accuracy": 0.9580988117573483,
+      "eval_religion_f1": 0.9574052840679498,
+      "eval_runtime": 10.9597,
+      "eval_safety_accuracy": 0.9221388367729831,
+      "eval_safety_f1": 0.9176523042532135,
+      "eval_samples_per_second": 875.387,
+      "eval_steps_per_second": 54.746,
+      "step": 9596
+    },
+    {
+      "epoch": 4.0016673614005835,
+      "grad_norm": 0.7530078887939453,
+      "learning_rate": 3.0638569604086846e-05,
+      "loss": 0.0852,
+      "step": 9600
+    },
+    {
+      "epoch": 4.043351396415173,
+      "grad_norm": 1.286962866783142,
+      "learning_rate": 3.042571306939123e-05,
+      "loss": 0.0617,
+      "step": 9700
+    },
+    {
+      "epoch": 4.0850354314297626,
+      "grad_norm": 1.2875369787216187,
+      "learning_rate": 3.0212856534695616e-05,
+      "loss": 0.0663,
+      "step": 9800
+    },
+    {
+      "epoch": 4.126719466444352,
+      "grad_norm": 2.4919862747192383,
+      "learning_rate": 3e-05,
+      "loss": 0.0609,
+      "step": 9900
+    },
+    {
+      "epoch": 4.168403501458942,
+      "grad_norm": 2.174107074737549,
+      "learning_rate": 2.9787143465304386e-05,
+      "loss": 0.0588,
+      "step": 10000
+    },
+    {
+      "epoch": 4.210087536473531,
+      "grad_norm": 1.7784463167190552,
+      "learning_rate": 2.9574286930608767e-05,
+      "loss": 0.0701,
+      "step": 10100
+    },
+    {
+      "epoch": 4.25177157148812,
+      "grad_norm": 1.0707236528396606,
+      "learning_rate": 2.936143039591316e-05,
+      "loss": 0.0738,
+      "step": 10200
+    },
+    {
+      "epoch": 4.293455606502709,
+      "grad_norm": 0.875899612903595,
+      "learning_rate": 2.9148573861217544e-05,
+      "loss": 0.0665,
+      "step": 10300
+    },
+    {
+      "epoch": 4.335139641517299,
+      "grad_norm": 2.555934190750122,
+      "learning_rate": 2.8935717326521926e-05,
+      "loss": 0.0639,
+      "step": 10400
+    },
+    {
+      "epoch": 4.376823676531888,
+      "grad_norm": 0.7635720372200012,
+      "learning_rate": 2.872286079182631e-05,
+      "loss": 0.0707,
+      "step": 10500
+    },
+    {
+      "epoch": 4.418507711546478,
+      "grad_norm": 1.4782235622406006,
+      "learning_rate": 2.8510004257130696e-05,
+      "loss": 0.0661,
+      "step": 10600
+    },
+    {
+      "epoch": 4.460191746561067,
+      "grad_norm": 2.679727077484131,
+      "learning_rate": 2.829714772243508e-05,
+      "loss": 0.0697,
+      "step": 10700
+    },
+    {
+      "epoch": 4.501875781575657,
+      "grad_norm": 0.8129181861877441,
+      "learning_rate": 2.8084291187739466e-05,
+      "loss": 0.0569,
+      "step": 10800
+    },
+    {
+      "epoch": 4.543559816590246,
+      "grad_norm": 1.4562526941299438,
+      "learning_rate": 2.7871434653043847e-05,
+      "loss": 0.0661,
+      "step": 10900
+    },
+    {
+      "epoch": 4.585243851604836,
+      "grad_norm": 1.9270412921905518,
+      "learning_rate": 2.7658578118348232e-05,
+      "loss": 0.057,
+      "step": 11000
+    },
+    {
+      "epoch": 4.626927886619425,
+      "grad_norm": 1.7193882465362549,
+      "learning_rate": 2.7445721583652617e-05,
+      "loss": 0.069,
+      "step": 11100
+    },
+    {
+      "epoch": 4.668611921634014,
+      "grad_norm": 2.909707546234131,
+      "learning_rate": 2.7232865048957006e-05,
+      "loss": 0.0639,
+      "step": 11200
+    },
+    {
+      "epoch": 4.710295956648603,
+      "grad_norm": 0.570263683795929,
+      "learning_rate": 2.702000851426139e-05,
+      "loss": 0.0654,
+      "step": 11300
+    },
+    {
+      "epoch": 4.751979991663193,
+      "grad_norm": 0.43538907170295715,
+      "learning_rate": 2.6807151979565776e-05,
+      "loss": 0.0633,
+      "step": 11400
+    },
+    {
+      "epoch": 4.793664026677782,
+      "grad_norm": 2.182251214981079,
+      "learning_rate": 2.659429544487016e-05,
+      "loss": 0.0697,
+      "step": 11500
+    },
+    {
+      "epoch": 4.835348061692372,
+      "grad_norm": 2.25455904006958,
+      "learning_rate": 2.6381438910174542e-05,
+      "loss": 0.0599,
+      "step": 11600
+    },
+    {
+      "epoch": 4.877032096706961,
+      "grad_norm": 1.227161169052124,
+      "learning_rate": 2.6168582375478927e-05,
+      "loss": 0.0637,
+      "step": 11700
+    },
+    {
+      "epoch": 4.918716131721551,
+      "grad_norm": 1.0838350057601929,
+      "learning_rate": 2.5955725840783312e-05,
+      "loss": 0.0635,
+      "step": 11800
+    },
+    {
+      "epoch": 4.96040016673614,
+      "grad_norm": 1.7880126237869263,
+      "learning_rate": 2.5742869306087697e-05,
+      "loss": 0.0573,
+      "step": 11900
+    },
+    {
+      "epoch": 5.0,
+      "eval_administration_accuracy": 0.9370439858244737,
+      "eval_administration_f1": 0.9329617589185005,
+      "eval_corruption_accuracy": 0.9603919116114238,
+      "eval_corruption_f1": 0.9582673518936824,
+      "eval_democracy_accuracy": 0.9550760892224307,
+      "eval_democracy_f1": 0.9548460506714369,
+      "eval_development_accuracy": 0.9215134459036898,
+      "eval_development_f1": 0.9164255722085631,
+      "eval_economy_accuracy": 0.931311236189285,
+      "eval_economy_f1": 0.9304486638126732,
+      "eval_education_accuracy": 0.9673754429851991,
+      "eval_education_f1": 0.9662633072587713,
+      "eval_environment_accuracy": 0.9785282468209298,
+      "eval_environment_f1": 0.9784438171410429,
+      "eval_instability_accuracy": 0.9468417761100688,
+      "eval_instability_f1": 0.9453606763183978,
+      "eval_leadership_accuracy": 0.8380237648530331,
+      "eval_leadership_f1": 0.8360259652707104,
+      "eval_loss": 0.2365979105234146,
+      "eval_overall_accuracy": 0.9393631436314364,
+      "eval_overall_f1": 0.9374392372783462,
+      "eval_race_accuracy": 0.9540337711069419,
+      "eval_race_f1": 0.9532943942985637,
+      "eval_religion_accuracy": 0.960287679799875,
+      "eval_religion_f1": 0.9582749968466362,
+      "eval_runtime": 10.946,
+      "eval_safety_accuracy": 0.9219303731498854,
+      "eval_safety_f1": 0.9186582927011756,
+      "eval_samples_per_second": 876.487,
+      "eval_steps_per_second": 54.815,
+      "step": 11995
+    }
+  ],
+  "logging_steps": 100,
+  "max_steps": 23990,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 2,
+        "early_stopping_threshold": 0.0
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 0
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 5.010440095766938e+16,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": {
+    "gradient_accumulation_steps": 1,
+    "learning_rate": 5e-05,
+    "num_train_epochs": 10
+  }
+}

run-1/checkpoint-11995/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4312a997194e916559618a59f3bbf4e1f98c127e89ae3b7fb5162b782863b12c
+size 5368

run-1/checkpoint-14394/config.json

@@ -0,0 +1,125 @@
+{
+  "architectures": [
+    "BertForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "classifier_dropout": null,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "LABEL_0",
+    "1": "LABEL_1",
+    "2": "LABEL_2",
+    "3": "LABEL_3",
+    "4": "LABEL_4",
+    "5": "LABEL_5",
+    "6": "LABEL_6",
+    "7": "LABEL_7",
+    "8": "LABEL_8",
+    "9": "LABEL_9",
+    "10": "LABEL_10",
+    "11": "LABEL_11",
+    "12": "LABEL_12",
+    "13": "LABEL_13",
+    "14": "LABEL_14",
+    "15": "LABEL_15",
+    "16": "LABEL_16",
+    "17": "LABEL_17",
+    "18": "LABEL_18",
+    "19": "LABEL_19",
+    "20": "LABEL_20",
+    "21": "LABEL_21",
+    "22": "LABEL_22",
+    "23": "LABEL_23",
+    "24": "LABEL_24",
+    "25": "LABEL_25",
+    "26": "LABEL_26",
+    "27": "LABEL_27",
+    "28": "LABEL_28",
+    "29": "LABEL_29",
+    "30": "LABEL_30",
+    "31": "LABEL_31",
+    "32": "LABEL_32",
+    "33": "LABEL_33",
+    "34": "LABEL_34",
+    "35": "LABEL_35",
+    "36": "LABEL_36",
+    "37": "LABEL_37",
+    "38": "LABEL_38",
+    "39": "LABEL_39",
+    "40": "LABEL_40",
+    "41": "LABEL_41",
+    "42": "LABEL_42",
+    "43": "LABEL_43",
+    "44": "LABEL_44",
+    "45": "LABEL_45",
+    "46": "LABEL_46",
+    "47": "LABEL_47"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "LABEL_0": 0,
+    "LABEL_1": 1,
+    "LABEL_10": 10,
+    "LABEL_11": 11,
+    "LABEL_12": 12,
+    "LABEL_13": 13,
+    "LABEL_14": 14,
+    "LABEL_15": 15,
+    "LABEL_16": 16,
+    "LABEL_17": 17,
+    "LABEL_18": 18,
+    "LABEL_19": 19,
+    "LABEL_2": 2,
+    "LABEL_20": 20,
+    "LABEL_21": 21,
+    "LABEL_22": 22,
+    "LABEL_23": 23,
+    "LABEL_24": 24,
+    "LABEL_25": 25,
+    "LABEL_26": 26,
+    "LABEL_27": 27,
+    "LABEL_28": 28,
+    "LABEL_29": 29,
+    "LABEL_3": 3,
+    "LABEL_30": 30,
+    "LABEL_31": 31,
+    "LABEL_32": 32,
+    "LABEL_33": 33,
+    "LABEL_34": 34,
+    "LABEL_35": 35,
+    "LABEL_36": 36,
+    "LABEL_37": 37,
+    "LABEL_38": 38,
+    "LABEL_39": 39,
+    "LABEL_4": 4,
+    "LABEL_40": 40,
+    "LABEL_41": 41,
+    "LABEL_42": 42,
+    "LABEL_43": 43,
+    "LABEL_44": 44,
+    "LABEL_45": 45,
+    "LABEL_46": 46,
+    "LABEL_47": 47,
+    "LABEL_5": 5,
+    "LABEL_6": 6,
+    "LABEL_7": 7,
+    "LABEL_8": 8,
+    "LABEL_9": 9
+  },
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "torch_dtype": "float32",
+  "transformers_version": "4.50.3",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 30522
+}

run-1/checkpoint-14394/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8824742ee74398fb20c851f3bb89873134e831281d1bbf422888e293f8935ae8
+size 438100144

run-1/checkpoint-14394/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5bb0c840d373af0be469db731bef1bbab09b8d8f40faedd6f7cc4b0fca8982b
+size 876321402

run-1/checkpoint-14394/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:8950df86f7a6bcf4605ec42c8c63f0262cb16b1bfc5127adb4efa5a0b1b1330d
+size 14244

run-1/checkpoint-14394/scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a6e7ffa8422af280d06886df2b30ff24e86e5d828f76e7ede0553e7b44d8672
+size 988

run-1/checkpoint-14394/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:5a3fd9e90bd54e92e68a90ddb2c5d56da08813d8d4ee1d58b8612f8c6dd37363
+size 1064

run-1/checkpoint-14394/trainer_state.json

@@ -0,0 +1,1252 @@
+{
+  "best_global_step": 11995,
+  "best_metric": 0.9374392372783462,
+  "best_model_checkpoint": "./results/run-1/checkpoint-11995",
+  "epoch": 6.0,
+  "eval_steps": 500,
+  "global_step": 14394,
+  "is_hyper_param_search": true,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.041684035014589414,
+      "grad_norm": 2.362386465072632,
+      "learning_rate": 9.900000000000002e-06,
+      "loss": 1.3944,
+      "step": 100
+    },
+    {
+      "epoch": 0.08336807002917883,
+      "grad_norm": 0.7483571767807007,
+      "learning_rate": 1.9900000000000003e-05,
+      "loss": 0.5576,
+      "step": 200
+    },
+    {
+      "epoch": 0.12505210504376824,
+      "grad_norm": 0.8563070297241211,
+      "learning_rate": 2.9900000000000002e-05,
+      "loss": 0.4502,
+      "step": 300
+    },
+    {
+      "epoch": 0.16673614005835766,
+      "grad_norm": 0.7200814485549927,
+      "learning_rate": 3.99e-05,
+      "loss": 0.388,
+      "step": 400
+    },
+    {
+      "epoch": 0.20842017507294705,
+      "grad_norm": 1.4553083181381226,
+      "learning_rate": 4.99e-05,
+      "loss": 0.3508,
+      "step": 500
+    },
+    {
+      "epoch": 0.25010421008753647,
+      "grad_norm": 0.9915399551391602,
+      "learning_rate": 4.978927203065135e-05,
+      "loss": 0.3186,
+      "step": 600
+    },
+    {
+      "epoch": 0.29178824510212586,
+      "grad_norm": 0.9978141188621521,
+      "learning_rate": 4.957641549595573e-05,
+      "loss": 0.3087,
+      "step": 700
+    },
+    {
+      "epoch": 0.3334722801167153,
+      "grad_norm": 1.3709461688995361,
+      "learning_rate": 4.936355896126011e-05,
+      "loss": 0.2976,
+      "step": 800
+    },
+    {
+      "epoch": 0.3751563151313047,
+      "grad_norm": 1.82745361328125,
+      "learning_rate": 4.91507024265645e-05,
+      "loss": 0.28,
+      "step": 900
+    },
+    {
+      "epoch": 0.4168403501458941,
+      "grad_norm": 1.2072925567626953,
+      "learning_rate": 4.893784589186888e-05,
+      "loss": 0.2751,
+      "step": 1000
+    },
+    {
+      "epoch": 0.45852438516048355,
+      "grad_norm": 1.0433937311172485,
+      "learning_rate": 4.872498935717327e-05,
+      "loss": 0.2712,
+      "step": 1100
+    },
+    {
+      "epoch": 0.5002084201750729,
+      "grad_norm": 1.1097549200057983,
+      "learning_rate": 4.851213282247765e-05,
+      "loss": 0.2909,
+      "step": 1200
+    },
+    {
+      "epoch": 0.5418924551896623,
+      "grad_norm": 1.2917771339416504,
+      "learning_rate": 4.8299276287782034e-05,
+      "loss": 0.2501,
+      "step": 1300
+    },
+    {
+      "epoch": 0.5835764902042517,
+      "grad_norm": 0.9812286496162415,
+      "learning_rate": 4.808641975308642e-05,
+      "loss": 0.2621,
+      "step": 1400
+    },
+    {
+      "epoch": 0.6252605252188412,
+      "grad_norm": 1.1786612272262573,
+      "learning_rate": 4.7873563218390804e-05,
+      "loss": 0.2435,
+      "step": 1500
+    },
+    {
+      "epoch": 0.6669445602334306,
+      "grad_norm": 1.4759145975112915,
+      "learning_rate": 4.766070668369519e-05,
+      "loss": 0.2398,
+      "step": 1600
+    },
+    {
+      "epoch": 0.70862859524802,
+      "grad_norm": 1.5548242330551147,
+      "learning_rate": 4.744785014899958e-05,
+      "loss": 0.2432,
+      "step": 1700
+    },
+    {
+      "epoch": 0.7503126302626094,
+      "grad_norm": 1.9005539417266846,
+      "learning_rate": 4.723499361430396e-05,
+      "loss": 0.2351,
+      "step": 1800
+    },
+    {
+      "epoch": 0.7919966652771988,
+      "grad_norm": 1.5831588506698608,
+      "learning_rate": 4.702213707960835e-05,
+      "loss": 0.2378,
+      "step": 1900
+    },
+    {
+      "epoch": 0.8336807002917882,
+      "grad_norm": 1.5482088327407837,
+      "learning_rate": 4.680928054491273e-05,
+      "loss": 0.2271,
+      "step": 2000
+    },
+    {
+      "epoch": 0.8753647353063777,
+      "grad_norm": 1.1437913179397583,
+      "learning_rate": 4.6596424010217114e-05,
+      "loss": 0.231,
+      "step": 2100
+    },
+    {
+      "epoch": 0.9170487703209671,
+      "grad_norm": 0.7985406517982483,
+      "learning_rate": 4.63835674755215e-05,
+      "loss": 0.228,
+      "step": 2200
+    },
+    {
+      "epoch": 0.9587328053355565,
+      "grad_norm": 1.9323142766952515,
+      "learning_rate": 4.6170710940825884e-05,
+      "loss": 0.2198,
+      "step": 2300
+    },
+    {
+      "epoch": 1.0,
+      "eval_administration_accuracy": 0.9322493224932249,
+      "eval_administration_f1": 0.9122911143651031,
+      "eval_corruption_accuracy": 0.9560141755263707,
+      "eval_corruption_f1": 0.9517304794081561,
+      "eval_democracy_accuracy": 0.9505941213258287,
+      "eval_democracy_f1": 0.9401838742047087,
+      "eval_development_accuracy": 0.9118198874296435,
+      "eval_development_f1": 0.9003553411809495,
+      "eval_economy_accuracy": 0.9264123410464874,
+      "eval_economy_f1": 0.9202465382357672,
+      "eval_education_accuracy": 0.9643527204502814,
+      "eval_education_f1": 0.9615509617901187,
+      "eval_environment_accuracy": 0.9799874921826142,
+      "eval_environment_f1": 0.9786142985651582,
+      "eval_instability_accuracy": 0.9417344173441734,
+      "eval_instability_f1": 0.9340921018628459,
+      "eval_leadership_accuracy": 0.8085261621846987,
+      "eval_leadership_f1": 0.7858417402549625,
+      "eval_loss": 0.21971678733825684,
+      "eval_overall_accuracy": 0.9324317281634354,
+      "eval_overall_f1": 0.9244250028140656,
+      "eval_race_accuracy": 0.9508025849489264,
+      "eval_race_f1": 0.9481714649378918,
+      "eval_religion_accuracy": 0.9530956848030019,
+      "eval_religion_f1": 0.9518106778632144,
+      "eval_runtime": 10.9084,
+      "eval_safety_accuracy": 0.9135918282259746,
+      "eval_safety_f1": 0.9082114410999095,
+      "eval_samples_per_second": 879.505,
+      "eval_steps_per_second": 55.003,
+      "step": 2399
+    },
+    {
+      "epoch": 1.0004168403501459,
+      "grad_norm": 1.2817295789718628,
+      "learning_rate": 4.5957854406130265e-05,
+      "loss": 0.2397,
+      "step": 2400
+    },
+    {
+      "epoch": 1.0421008753647354,
+      "grad_norm": 1.3623204231262207,
+      "learning_rate": 4.5744997871434654e-05,
+      "loss": 0.1885,
+      "step": 2500
+    },
+    {
+      "epoch": 1.0837849103793247,
+      "grad_norm": 1.2534806728363037,
+      "learning_rate": 4.553214133673904e-05,
+      "loss": 0.1843,
+      "step": 2600
+    },
+    {
+      "epoch": 1.1254689453939142,
+      "grad_norm": 1.2860488891601562,
+      "learning_rate": 4.5319284802043424e-05,
+      "loss": 0.1817,
+      "step": 2700
+    },
+    {
+      "epoch": 1.1671529804085035,
+      "grad_norm": 1.2887414693832397,
+      "learning_rate": 4.510642826734781e-05,
+      "loss": 0.1798,
+      "step": 2800
+    },
+    {
+      "epoch": 1.208837015423093,
+      "grad_norm": 1.973421335220337,
+      "learning_rate": 4.4893571732652194e-05,
+      "loss": 0.1878,
+      "step": 2900
+    },
+    {
+      "epoch": 1.2505210504376825,
+      "grad_norm": 0.7910940051078796,
+      "learning_rate": 4.468071519795658e-05,
+      "loss": 0.182,
+      "step": 3000
+    },
+    {
+      "epoch": 1.2922050854522718,
+      "grad_norm": 1.7422680854797363,
+      "learning_rate": 4.4467858663260964e-05,
+      "loss": 0.1833,
+      "step": 3100
+    },
+    {
+      "epoch": 1.3338891204668613,
+      "grad_norm": 1.1493451595306396,
+      "learning_rate": 4.4255002128565345e-05,
+      "loss": 0.1766,
+      "step": 3200
+    },
+    {
+      "epoch": 1.3755731554814505,
+      "grad_norm": 2.179731845855713,
+      "learning_rate": 4.4042145593869734e-05,
+      "loss": 0.1785,
+      "step": 3300
+    },
+    {
+      "epoch": 1.41725719049604,
+      "grad_norm": 1.5773117542266846,
+      "learning_rate": 4.3829289059174115e-05,
+      "loss": 0.1901,
+      "step": 3400
+    },
+    {
+      "epoch": 1.4589412255106295,
+      "grad_norm": 2.4681854248046875,
+      "learning_rate": 4.3616432524478504e-05,
+      "loss": 0.1833,
+      "step": 3500
+    },
+    {
+      "epoch": 1.5006252605252188,
+      "grad_norm": 2.391803503036499,
+      "learning_rate": 4.340357598978289e-05,
+      "loss": 0.1677,
+      "step": 3600
+    },
+    {
+      "epoch": 1.5423092955398081,
+      "grad_norm": 0.8182645440101624,
+      "learning_rate": 4.3190719455087274e-05,
+      "loss": 0.1666,
+      "step": 3700
+    },
+    {
+      "epoch": 1.5839933305543976,
+      "grad_norm": 1.2136952877044678,
+      "learning_rate": 4.297786292039166e-05,
+      "loss": 0.1787,
+      "step": 3800
+    },
+    {
+      "epoch": 1.6256773655689871,
+      "grad_norm": 2.435816764831543,
+      "learning_rate": 4.2765006385696044e-05,
+      "loss": 0.1678,
+      "step": 3900
+    },
+    {
+      "epoch": 1.6673614005835766,
+      "grad_norm": 1.5172932147979736,
+      "learning_rate": 4.2552149851000425e-05,
+      "loss": 0.1779,
+      "step": 4000
+    },
+    {
+      "epoch": 1.709045435598166,
+      "grad_norm": 1.4677119255065918,
+      "learning_rate": 4.2339293316304814e-05,
+      "loss": 0.1824,
+      "step": 4100
+    },
+    {
+      "epoch": 1.7507294706127552,
+      "grad_norm": 1.8327163457870483,
+      "learning_rate": 4.2126436781609195e-05,
+      "loss": 0.1719,
+      "step": 4200
+    },
+    {
+      "epoch": 1.7924135056273447,
+      "grad_norm": 1.2324726581573486,
+      "learning_rate": 4.1913580246913584e-05,
+      "loss": 0.1592,
+      "step": 4300
+    },
+    {
+      "epoch": 1.8340975406419342,
+      "grad_norm": 3.6276469230651855,
+      "learning_rate": 4.1700723712217965e-05,
+      "loss": 0.1636,
+      "step": 4400
+    },
+    {
+      "epoch": 1.8757815756565237,
+      "grad_norm": 0.940485417842865,
+      "learning_rate": 4.148786717752235e-05,
+      "loss": 0.1639,
+      "step": 4500
+    },
+    {
+      "epoch": 1.917465610671113,
+      "grad_norm": 2.002577781677246,
+      "learning_rate": 4.127501064282674e-05,
+      "loss": 0.1756,
+      "step": 4600
+    },
+    {
+      "epoch": 1.9591496456857023,
+      "grad_norm": 2.6062567234039307,
+      "learning_rate": 4.1062154108131124e-05,
+      "loss": 0.1729,
+      "step": 4700
+    },
+    {
+      "epoch": 2.0,
+      "eval_administration_accuracy": 0.9374609130706691,
+      "eval_administration_f1": 0.9302102302828306,
+      "eval_corruption_accuracy": 0.9603919116114238,
+      "eval_corruption_f1": 0.9585025895331236,
+      "eval_democracy_accuracy": 0.9562226391494684,
+      "eval_democracy_f1": 0.9505005443286667,
+      "eval_development_accuracy": 0.9199499687304565,
+      "eval_development_f1": 0.9145013995891959,
+      "eval_economy_accuracy": 0.9319366270585783,
+      "eval_economy_f1": 0.9302256375627648,
+      "eval_education_accuracy": 0.9651865749426725,
+      "eval_education_f1": 0.9641689686262792,
+      "eval_environment_accuracy": 0.9806128830519074,
+      "eval_environment_f1": 0.9798389855050841,
+      "eval_instability_accuracy": 0.9244319366270586,
+      "eval_instability_f1": 0.9301250908302109,
+      "eval_leadership_accuracy": 0.8395872420262664,
+      "eval_leadership_f1": 0.830858580138579,
+      "eval_loss": 0.20445191860198975,
+      "eval_overall_accuracy": 0.9366791744840525,
+      "eval_overall_f1": 0.9342598776015526,
+      "eval_race_accuracy": 0.9545549301646863,
+      "eval_race_f1": 0.9531323667533257,
+      "eval_religion_accuracy": 0.9575776526996039,
+      "eval_religion_f1": 0.9560680494194458,
+      "eval_runtime": 10.9745,
+      "eval_safety_accuracy": 0.9122368146758391,
+      "eval_safety_f1": 0.9129860886491253,
+      "eval_samples_per_second": 874.208,
+      "eval_steps_per_second": 54.672,
+      "step": 4798
+    },
+    {
+      "epoch": 2.0008336807002918,
+      "grad_norm": 1.0561970472335815,
+      "learning_rate": 4.085142613878246e-05,
+      "loss": 0.174,
+      "step": 4800
+    },
+    {
+      "epoch": 2.0425177157148813,
+      "grad_norm": 0.9414407014846802,
+      "learning_rate": 4.0638569604086845e-05,
+      "loss": 0.1259,
+      "step": 4900
+    },
+    {
+      "epoch": 2.084201750729471,
+      "grad_norm": 1.0318489074707031,
+      "learning_rate": 4.042571306939123e-05,
+      "loss": 0.1258,
+      "step": 5000
+    },
+    {
+      "epoch": 2.12588578574406,
+      "grad_norm": 2.4367470741271973,
+      "learning_rate": 4.0212856534695615e-05,
+      "loss": 0.1186,
+      "step": 5100
+    },
+    {
+      "epoch": 2.1675698207586493,
+      "grad_norm": 1.8129918575286865,
+      "learning_rate": 4e-05,
+      "loss": 0.1243,
+      "step": 5200
+    },
+    {
+      "epoch": 2.209253855773239,
+      "grad_norm": 2.235281467437744,
+      "learning_rate": 3.978714346530439e-05,
+      "loss": 0.1169,
+      "step": 5300
+    },
+    {
+      "epoch": 2.2509378907878284,
+      "grad_norm": 3.7830569744110107,
+      "learning_rate": 3.957428693060877e-05,
+      "loss": 0.1177,
+      "step": 5400
+    },
+    {
+      "epoch": 2.292621925802418,
+      "grad_norm": 1.2535426616668701,
+      "learning_rate": 3.9361430395913155e-05,
+      "loss": 0.1204,
+      "step": 5500
+    },
+    {
+      "epoch": 2.334305960817007,
+      "grad_norm": 1.443596601486206,
+      "learning_rate": 3.914857386121754e-05,
+      "loss": 0.1187,
+      "step": 5600
+    },
+    {
+      "epoch": 2.3759899958315964,
+      "grad_norm": 1.0241508483886719,
+      "learning_rate": 3.8935717326521925e-05,
+      "loss": 0.1222,
+      "step": 5700
+    },
+    {
+      "epoch": 2.417674030846186,
+      "grad_norm": 1.923004150390625,
+      "learning_rate": 3.8722860791826307e-05,
+      "loss": 0.1246,
+      "step": 5800
+    },
+    {
+      "epoch": 2.4593580658607754,
+      "grad_norm": 0.7686344385147095,
+      "learning_rate": 3.8510004257130695e-05,
+      "loss": 0.13,
+      "step": 5900
+    },
+    {
+      "epoch": 2.501042100875365,
+      "grad_norm": 1.762736439704895,
+      "learning_rate": 3.8297147722435077e-05,
+      "loss": 0.1263,
+      "step": 6000
+    },
+    {
+      "epoch": 2.542726135889954,
+      "grad_norm": 1.9006534814834595,
+      "learning_rate": 3.8084291187739465e-05,
+      "loss": 0.1152,
+      "step": 6100
+    },
+    {
+      "epoch": 2.5844101709045435,
+      "grad_norm": 2.003476858139038,
+      "learning_rate": 3.787143465304385e-05,
+      "loss": 0.1318,
+      "step": 6200
+    },
+    {
+      "epoch": 2.626094205919133,
+      "grad_norm": 1.8864182233810425,
+      "learning_rate": 3.7658578118348235e-05,
+      "loss": 0.1296,
+      "step": 6300
+    },
+    {
+      "epoch": 2.6677782409337225,
+      "grad_norm": 1.567968726158142,
+      "learning_rate": 3.744572158365262e-05,
+      "loss": 0.1292,
+      "step": 6400
+    },
+    {
+      "epoch": 2.7094622759483116,
+      "grad_norm": 1.7688707113265991,
+      "learning_rate": 3.7232865048957005e-05,
+      "loss": 0.1258,
+      "step": 6500
+    },
+    {
+      "epoch": 2.751146310962901,
+      "grad_norm": 2.138125419616699,
+      "learning_rate": 3.7020008514261387e-05,
+      "loss": 0.1208,
+      "step": 6600
+    },
+    {
+      "epoch": 2.7928303459774906,
+      "grad_norm": 3.3148305416107178,
+      "learning_rate": 3.6807151979565775e-05,
+      "loss": 0.1253,
+      "step": 6700
+    },
+    {
+      "epoch": 2.83451438099208,
+      "grad_norm": 1.8787152767181396,
+      "learning_rate": 3.6594295444870157e-05,
+      "loss": 0.1161,
+      "step": 6800
+    },
+    {
+      "epoch": 2.8761984160066696,
+      "grad_norm": 0.7995413541793823,
+      "learning_rate": 3.6381438910174545e-05,
+      "loss": 0.1196,
+      "step": 6900
+    },
+    {
+      "epoch": 2.917882451021259,
+      "grad_norm": 1.973639965057373,
+      "learning_rate": 3.6168582375478927e-05,
+      "loss": 0.1227,
+      "step": 7000
+    },
+    {
+      "epoch": 2.959566486035848,
+      "grad_norm": 1.0217853784561157,
+      "learning_rate": 3.5955725840783315e-05,
+      "loss": 0.125,
+      "step": 7100
+    },
+    {
+      "epoch": 3.0,
+      "eval_administration_accuracy": 0.9373566812591203,
+      "eval_administration_f1": 0.9333482432043958,
+      "eval_corruption_accuracy": 0.9609130706691682,
+      "eval_corruption_f1": 0.958848962871965,
+      "eval_democracy_accuracy": 0.9567437982072129,
+      "eval_democracy_f1": 0.9532298494465137,
+      "eval_development_accuracy": 0.9128622055451324,
+      "eval_development_f1": 0.9113351170841872,
+      "eval_economy_accuracy": 0.9317281634354805,
+      "eval_economy_f1": 0.9292396676492435,
+      "eval_education_accuracy": 0.965603502188868,
+      "eval_education_f1": 0.9637981071832747,
+      "eval_environment_accuracy": 0.9804044194288096,
+      "eval_environment_f1": 0.9788214035384287,
+      "eval_instability_accuracy": 0.9490306441525954,
+      "eval_instability_f1": 0.9469525951807101,
+      "eval_leadership_accuracy": 0.8446946007921617,
+      "eval_leadership_f1": 0.840338137042441,
+      "eval_loss": 0.20486579835414886,
+      "eval_overall_accuracy": 0.9398235007991106,
+      "eval_overall_f1": 0.9372835834491008,
+      "eval_race_accuracy": 0.9540337711069419,
+      "eval_race_f1": 0.9520143591935892,
+      "eval_religion_accuracy": 0.959349593495935,
+      "eval_religion_f1": 0.9575545048864775,
+      "eval_runtime": 10.9937,
+      "eval_safety_accuracy": 0.9251615593079008,
+      "eval_safety_f1": 0.9219220541079818,
+      "eval_samples_per_second": 872.682,
+      "eval_steps_per_second": 54.577,
+      "step": 7197
+    },
+    {
+      "epoch": 3.0012505210504377,
+      "grad_norm": 0.8939034938812256,
+      "learning_rate": 3.5744997871434655e-05,
+      "loss": 0.1224,
+      "step": 7200
+    },
+    {
+      "epoch": 3.042934556065027,
+      "grad_norm": 2.387557029724121,
+      "learning_rate": 3.5532141336739036e-05,
+      "loss": 0.0924,
+      "step": 7300
+    },
+    {
+      "epoch": 3.0846185910796167,
+      "grad_norm": 0.7587341666221619,
+      "learning_rate": 3.5319284802043425e-05,
+      "loss": 0.0838,
+      "step": 7400
+    },
+    {
+      "epoch": 3.1263026260942057,
+      "grad_norm": 2.06026291847229,
+      "learning_rate": 3.5106428267347806e-05,
+      "loss": 0.0894,
+      "step": 7500
+    },
+    {
+      "epoch": 3.1679866611087952,
+      "grad_norm": 2.001525402069092,
+      "learning_rate": 3.4893571732652195e-05,
+      "loss": 0.0913,
+      "step": 7600
+    },
+    {
+      "epoch": 3.2096706961233847,
+      "grad_norm": 2.0007410049438477,
+      "learning_rate": 3.468071519795658e-05,
+      "loss": 0.0891,
+      "step": 7700
+    },
+    {
+      "epoch": 3.2513547311379742,
+      "grad_norm": 1.6416945457458496,
+      "learning_rate": 3.4467858663260965e-05,
+      "loss": 0.0919,
+      "step": 7800
+    },
+    {
+      "epoch": 3.2930387661525637,
+      "grad_norm": 1.653064489364624,
+      "learning_rate": 3.425500212856535e-05,
+      "loss": 0.0932,
+      "step": 7900
+    },
+    {
+      "epoch": 3.334722801167153,
+      "grad_norm": 1.7365379333496094,
+      "learning_rate": 3.4042145593869735e-05,
+      "loss": 0.0898,
+      "step": 8000
+    },
+    {
+      "epoch": 3.3764068361817423,
+      "grad_norm": 1.561340093612671,
+      "learning_rate": 3.3829289059174116e-05,
+      "loss": 0.0866,
+      "step": 8100
+    },
+    {
+      "epoch": 3.418090871196332,
+      "grad_norm": 0.7200934886932373,
+      "learning_rate": 3.3616432524478505e-05,
+      "loss": 0.0844,
+      "step": 8200
+    },
+    {
+      "epoch": 3.4597749062109213,
+      "grad_norm": 0.8205280303955078,
+      "learning_rate": 3.3405704555129844e-05,
+      "loss": 0.0914,
+      "step": 8300
+    },
+    {
+      "epoch": 3.5014589412255104,
+      "grad_norm": 1.7642834186553955,
+      "learning_rate": 3.319284802043423e-05,
+      "loss": 0.0873,
+      "step": 8400
+    },
+    {
+      "epoch": 3.5431429762401,
+      "grad_norm": 2.3324527740478516,
+      "learning_rate": 3.2979991485738614e-05,
+      "loss": 0.0808,
+      "step": 8500
+    },
+    {
+      "epoch": 3.5848270112546894,
+      "grad_norm": 1.7013909816741943,
+      "learning_rate": 3.2767134951043e-05,
+      "loss": 0.094,
+      "step": 8600
+    },
+    {
+      "epoch": 3.626511046269279,
+      "grad_norm": 4.724839210510254,
+      "learning_rate": 3.2554278416347384e-05,
+      "loss": 0.0878,
+      "step": 8700
+    },
+    {
+      "epoch": 3.6681950812838684,
+      "grad_norm": 3.1913349628448486,
+      "learning_rate": 3.2341421881651766e-05,
+      "loss": 0.0918,
+      "step": 8800
+    },
+    {
+      "epoch": 3.709879116298458,
+      "grad_norm": 2.2082605361938477,
+      "learning_rate": 3.2128565346956154e-05,
+      "loss": 0.0846,
+      "step": 8900
+    },
+    {
+      "epoch": 3.751563151313047,
+      "grad_norm": 2.2428789138793945,
+      "learning_rate": 3.1915708812260536e-05,
+      "loss": 0.0878,
+      "step": 9000
+    },
+    {
+      "epoch": 3.7932471863276365,
+      "grad_norm": 2.0431671142578125,
+      "learning_rate": 3.170285227756492e-05,
+      "loss": 0.0964,
+      "step": 9100
+    },
+    {
+      "epoch": 3.834931221342226,
+      "grad_norm": 1.9629932641983032,
+      "learning_rate": 3.1489995742869306e-05,
+      "loss": 0.0969,
+      "step": 9200
+    },
+    {
+      "epoch": 3.8766152563568155,
+      "grad_norm": 1.0876961946487427,
+      "learning_rate": 3.1277139208173694e-05,
+      "loss": 0.0864,
+      "step": 9300
+    },
+    {
+      "epoch": 3.9182992913714045,
+      "grad_norm": 2.003713607788086,
+      "learning_rate": 3.1064282673478076e-05,
+      "loss": 0.0843,
+      "step": 9400
+    },
+    {
+      "epoch": 3.959983326385994,
+      "grad_norm": 1.3784735202789307,
+      "learning_rate": 3.0851426138782464e-05,
+      "loss": 0.0866,
+      "step": 9500
+    },
+    {
+      "epoch": 4.0,
+      "eval_administration_accuracy": 0.9397540129247446,
+      "eval_administration_f1": 0.9326250155599194,
+      "eval_corruption_accuracy": 0.9600792161767772,
+      "eval_corruption_f1": 0.9585031567143241,
+      "eval_democracy_accuracy": 0.9542422347300397,
+      "eval_democracy_f1": 0.9522572809405702,
+      "eval_development_accuracy": 0.9216176777152387,
+      "eval_development_f1": 0.9176939504856475,
+      "eval_economy_accuracy": 0.9336043360433605,
+      "eval_economy_f1": 0.9317409802418596,
+      "eval_education_accuracy": 0.968313529289139,
+      "eval_education_f1": 0.9659804103888047,
+      "eval_environment_accuracy": 0.9789451740671253,
+      "eval_environment_f1": 0.9783634505314646,
+      "eval_instability_accuracy": 0.9435063581405044,
+      "eval_instability_f1": 0.9430673845492272,
+      "eval_leadership_accuracy": 0.8415676464456953,
+      "eval_leadership_f1": 0.8394569407264164,
+      "eval_loss": 0.22299017012119293,
+      "eval_overall_accuracy": 0.9397540129247446,
+      "eval_overall_f1": 0.9374286167544477,
+      "eval_race_accuracy": 0.9551803210339795,
+      "eval_race_f1": 0.9543972425939777,
+      "eval_religion_accuracy": 0.9580988117573483,
+      "eval_religion_f1": 0.9574052840679498,
+      "eval_runtime": 10.9597,
+      "eval_safety_accuracy": 0.9221388367729831,
+      "eval_safety_f1": 0.9176523042532135,
+      "eval_samples_per_second": 875.387,
+      "eval_steps_per_second": 54.746,
+      "step": 9596
+    },
+    {
+      "epoch": 4.0016673614005835,
+      "grad_norm": 0.7530078887939453,
+      "learning_rate": 3.0638569604086846e-05,
+      "loss": 0.0852,
+      "step": 9600
+    },
+    {
+      "epoch": 4.043351396415173,
+      "grad_norm": 1.286962866783142,
+      "learning_rate": 3.042571306939123e-05,
+      "loss": 0.0617,
+      "step": 9700
+    },
+    {
+      "epoch": 4.0850354314297626,
+      "grad_norm": 1.2875369787216187,
+      "learning_rate": 3.0212856534695616e-05,
+      "loss": 0.0663,
+      "step": 9800
+    },
+    {
+      "epoch": 4.126719466444352,
+      "grad_norm": 2.4919862747192383,
+      "learning_rate": 3e-05,
+      "loss": 0.0609,
+      "step": 9900
+    },
+    {
+      "epoch": 4.168403501458942,
+      "grad_norm": 2.174107074737549,
+      "learning_rate": 2.9787143465304386e-05,
+      "loss": 0.0588,
+      "step": 10000
+    },
+    {
+      "epoch": 4.210087536473531,
+      "grad_norm": 1.7784463167190552,
+      "learning_rate": 2.9574286930608767e-05,
+      "loss": 0.0701,
+      "step": 10100
+    },
+    {
+      "epoch": 4.25177157148812,
+      "grad_norm": 1.0707236528396606,
+      "learning_rate": 2.936143039591316e-05,
+      "loss": 0.0738,
+      "step": 10200
+    },
+    {
+      "epoch": 4.293455606502709,
+      "grad_norm": 0.875899612903595,
+      "learning_rate": 2.9148573861217544e-05,
+      "loss": 0.0665,
+      "step": 10300
+    },
+    {
+      "epoch": 4.335139641517299,
+      "grad_norm": 2.555934190750122,
+      "learning_rate": 2.8935717326521926e-05,
+      "loss": 0.0639,
+      "step": 10400
+    },
+    {
+      "epoch": 4.376823676531888,
+      "grad_norm": 0.7635720372200012,
+      "learning_rate": 2.872286079182631e-05,
+      "loss": 0.0707,
+      "step": 10500
+    },
+    {
+      "epoch": 4.418507711546478,
+      "grad_norm": 1.4782235622406006,
+      "learning_rate": 2.8510004257130696e-05,
+      "loss": 0.0661,
+      "step": 10600
+    },
+    {
+      "epoch": 4.460191746561067,
+      "grad_norm": 2.679727077484131,
+      "learning_rate": 2.829714772243508e-05,
+      "loss": 0.0697,
+      "step": 10700
+    },
+    {
+      "epoch": 4.501875781575657,
+      "grad_norm": 0.8129181861877441,
+      "learning_rate": 2.8084291187739466e-05,
+      "loss": 0.0569,
+      "step": 10800
+    },
+    {
+      "epoch": 4.543559816590246,
+      "grad_norm": 1.4562526941299438,
+      "learning_rate": 2.7871434653043847e-05,
+      "loss": 0.0661,
+      "step": 10900
+    },
+    {
+      "epoch": 4.585243851604836,
+      "grad_norm": 1.9270412921905518,
+      "learning_rate": 2.7658578118348232e-05,
+      "loss": 0.057,
+      "step": 11000
+    },
+    {
+      "epoch": 4.626927886619425,
+      "grad_norm": 1.7193882465362549,
+      "learning_rate": 2.7445721583652617e-05,
+      "loss": 0.069,
+      "step": 11100
+    },
+    {
+      "epoch": 4.668611921634014,
+      "grad_norm": 2.909707546234131,
+      "learning_rate": 2.7232865048957006e-05,
+      "loss": 0.0639,
+      "step": 11200
+    },
+    {
+      "epoch": 4.710295956648603,
+      "grad_norm": 0.570263683795929,
+      "learning_rate": 2.702000851426139e-05,
+      "loss": 0.0654,
+      "step": 11300
+    },
+    {
+      "epoch": 4.751979991663193,
+      "grad_norm": 0.43538907170295715,
+      "learning_rate": 2.6807151979565776e-05,
+      "loss": 0.0633,
+      "step": 11400
+    },
+    {
+      "epoch": 4.793664026677782,
+      "grad_norm": 2.182251214981079,
+      "learning_rate": 2.659429544487016e-05,
+      "loss": 0.0697,
+      "step": 11500
+    },
+    {
+      "epoch": 4.835348061692372,
+      "grad_norm": 2.25455904006958,
+      "learning_rate": 2.6381438910174542e-05,
+      "loss": 0.0599,
+      "step": 11600
+    },
+    {
+      "epoch": 4.877032096706961,
+      "grad_norm": 1.227161169052124,
+      "learning_rate": 2.6168582375478927e-05,
+      "loss": 0.0637,
+      "step": 11700
+    },
+    {
+      "epoch": 4.918716131721551,
+      "grad_norm": 1.0838350057601929,
+      "learning_rate": 2.5955725840783312e-05,
+      "loss": 0.0635,
+      "step": 11800
+    },
+    {
+      "epoch": 4.96040016673614,
+      "grad_norm": 1.7880126237869263,
+      "learning_rate": 2.5742869306087697e-05,
+      "loss": 0.0573,
+      "step": 11900
+    },
+    {
+      "epoch": 5.0,
+      "eval_administration_accuracy": 0.9370439858244737,
+      "eval_administration_f1": 0.9329617589185005,
+      "eval_corruption_accuracy": 0.9603919116114238,
+      "eval_corruption_f1": 0.9582673518936824,
+      "eval_democracy_accuracy": 0.9550760892224307,
+      "eval_democracy_f1": 0.9548460506714369,
+      "eval_development_accuracy": 0.9215134459036898,
+      "eval_development_f1": 0.9164255722085631,
+      "eval_economy_accuracy": 0.931311236189285,
+      "eval_economy_f1": 0.9304486638126732,
+      "eval_education_accuracy": 0.9673754429851991,
+      "eval_education_f1": 0.9662633072587713,
+      "eval_environment_accuracy": 0.9785282468209298,
+      "eval_environment_f1": 0.9784438171410429,
+      "eval_instability_accuracy": 0.9468417761100688,
+      "eval_instability_f1": 0.9453606763183978,
+      "eval_leadership_accuracy": 0.8380237648530331,
+      "eval_leadership_f1": 0.8360259652707104,
+      "eval_loss": 0.2365979105234146,
+      "eval_overall_accuracy": 0.9393631436314364,
+      "eval_overall_f1": 0.9374392372783462,
+      "eval_race_accuracy": 0.9540337711069419,
+      "eval_race_f1": 0.9532943942985637,
+      "eval_religion_accuracy": 0.960287679799875,
+      "eval_religion_f1": 0.9582749968466362,
+      "eval_runtime": 10.946,
+      "eval_safety_accuracy": 0.9219303731498854,
+      "eval_safety_f1": 0.9186582927011756,
+      "eval_samples_per_second": 876.487,
+      "eval_steps_per_second": 54.815,
+      "step": 11995
+    },
+    {
+      "epoch": 5.00208420175073,
+      "grad_norm": 1.8319880962371826,
+      "learning_rate": 2.5530012771392082e-05,
+      "loss": 0.0644,
+      "step": 12000
+    },
+    {
+      "epoch": 5.0437682367653185,
+      "grad_norm": 0.9142705798149109,
+      "learning_rate": 2.5317156236696464e-05,
+      "loss": 0.0468,
+      "step": 12100
+    },
+    {
+      "epoch": 5.085452271779908,
+      "grad_norm": 1.5496271848678589,
+      "learning_rate": 2.5104299702000856e-05,
+      "loss": 0.0513,
+      "step": 12200
+    },
+    {
+      "epoch": 5.1271363067944975,
+      "grad_norm": 1.026939034461975,
+      "learning_rate": 2.4891443167305237e-05,
+      "loss": 0.0505,
+      "step": 12300
+    },
+    {
+      "epoch": 5.168820341809087,
+      "grad_norm": 0.9045308232307434,
+      "learning_rate": 2.4678586632609622e-05,
+      "loss": 0.0423,
+      "step": 12400
+    },
+    {
+      "epoch": 5.2105043768236765,
+      "grad_norm": 1.3580307960510254,
+      "learning_rate": 2.4465730097914007e-05,
+      "loss": 0.044,
+      "step": 12500
+    },
+    {
+      "epoch": 5.252188411838266,
+      "grad_norm": 2.3576512336730957,
+      "learning_rate": 2.4252873563218392e-05,
+      "loss": 0.0495,
+      "step": 12600
+    },
+    {
+      "epoch": 5.2938724468528555,
+      "grad_norm": 0.727289080619812,
+      "learning_rate": 2.4040017028522777e-05,
+      "loss": 0.0501,
+      "step": 12700
+    },
+    {
+      "epoch": 5.335556481867445,
+      "grad_norm": 0.7748624086380005,
+      "learning_rate": 2.3827160493827162e-05,
+      "loss": 0.0479,
+      "step": 12800
+    },
+    {
+      "epoch": 5.3772405168820345,
+      "grad_norm": 1.302977204322815,
+      "learning_rate": 2.3614303959131544e-05,
+      "loss": 0.046,
+      "step": 12900
+    },
+    {
+      "epoch": 5.418924551896623,
+      "grad_norm": 1.363124132156372,
+      "learning_rate": 2.3401447424435932e-05,
+      "loss": 0.0432,
+      "step": 13000
+    },
+    {
+      "epoch": 5.460608586911213,
+      "grad_norm": 0.7924700379371643,
+      "learning_rate": 2.3188590889740317e-05,
+      "loss": 0.0458,
+      "step": 13100
+    },
+    {
+      "epoch": 5.502292621925802,
+      "grad_norm": 0.6111019849777222,
+      "learning_rate": 2.29757343550447e-05,
+      "loss": 0.0466,
+      "step": 13200
+    },
+    {
+      "epoch": 5.543976656940392,
+      "grad_norm": 1.055407166481018,
+      "learning_rate": 2.2762877820349084e-05,
+      "loss": 0.0487,
+      "step": 13300
+    },
+    {
+      "epoch": 5.585660691954981,
+      "grad_norm": 0.47858452796936035,
+      "learning_rate": 2.255002128565347e-05,
+      "loss": 0.0448,
+      "step": 13400
+    },
+    {
+      "epoch": 5.627344726969571,
+      "grad_norm": 2.17866587638855,
+      "learning_rate": 2.2339293316304812e-05,
+      "loss": 0.0454,
+      "step": 13500
+    },
+    {
+      "epoch": 5.66902876198416,
+      "grad_norm": 0.9695106744766235,
+      "learning_rate": 2.2126436781609197e-05,
+      "loss": 0.0487,
+      "step": 13600
+    },
+    {
+      "epoch": 5.71071279699875,
+      "grad_norm": 1.223860502243042,
+      "learning_rate": 2.1913580246913582e-05,
+      "loss": 0.0464,
+      "step": 13700
+    },
+    {
+      "epoch": 5.752396832013339,
+      "grad_norm": 0.8663516640663147,
+      "learning_rate": 2.1700723712217967e-05,
+      "loss": 0.044,
+      "step": 13800
+    },
+    {
+      "epoch": 5.794080867027929,
+      "grad_norm": 0.7240914106369019,
+      "learning_rate": 2.148786717752235e-05,
+      "loss": 0.0463,
+      "step": 13900
+    },
+    {
+      "epoch": 5.835764902042518,
+      "grad_norm": 1.0169471502304077,
+      "learning_rate": 2.1275010642826737e-05,
+      "loss": 0.0449,
+      "step": 14000
+    },
+    {
+      "epoch": 5.877448937057107,
+      "grad_norm": 2.1470980644226074,
+      "learning_rate": 2.1062154108131122e-05,
+      "loss": 0.0387,
+      "step": 14100
+    },
+    {
+      "epoch": 5.919132972071696,
+      "grad_norm": 0.8170746564865112,
+      "learning_rate": 2.0849297573435507e-05,
+      "loss": 0.0444,
+      "step": 14200
+    },
+    {
+      "epoch": 5.960817007086286,
+      "grad_norm": 1.886773943901062,
+      "learning_rate": 2.063644103873989e-05,
+      "loss": 0.0408,
+      "step": 14300
+    },
+    {
+      "epoch": 6.0,
+      "eval_administration_accuracy": 0.9392328538670002,
+      "eval_administration_f1": 0.9335586868984106,
+      "eval_corruption_accuracy": 0.9608088388576194,
+      "eval_corruption_f1": 0.9586222965202317,
+      "eval_democracy_accuracy": 0.9575776526996039,
+      "eval_democracy_f1": 0.9552908906267366,
+      "eval_development_accuracy": 0.9196372732958099,
+      "eval_development_f1": 0.9170852123784591,
+      "eval_economy_accuracy": 0.9321450906816761,
+      "eval_economy_f1": 0.9289418561764905,
+      "eval_education_accuracy": 0.9687304565353346,
+      "eval_education_f1": 0.9666551400031382,
+      "eval_environment_accuracy": 0.9798832603710652,
+      "eval_environment_f1": 0.9786999163300606,
+      "eval_instability_accuracy": 0.9454867625599332,
+      "eval_instability_f1": 0.9422890312692406,
+      "eval_leadership_accuracy": 0.8316656243485512,
+      "eval_leadership_f1": 0.833450633159851,
+      "eval_loss": 0.2633351981639862,
+      "eval_overall_accuracy": 0.9391546800083385,
+      "eval_overall_f1": 0.937151862792397,
+      "eval_race_accuracy": 0.9554930164686263,
+      "eval_race_f1": 0.9546556151565829,
+      "eval_religion_accuracy": 0.9575776526996039,
+      "eval_religion_f1": 0.9567690967660134,
+      "eval_runtime": 11.0311,
+      "eval_safety_accuracy": 0.9216176777152387,
+      "eval_safety_f1": 0.9198039782235481,
+      "eval_samples_per_second": 869.723,
+      "eval_steps_per_second": 54.392,
+      "step": 14394
+    }
+  ],
+  "logging_steps": 100,
+  "max_steps": 23990,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 2,
+        "early_stopping_threshold": 0.0
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 1
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 6.020887973363712e+16,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": {
+    "gradient_accumulation_steps": 1,
+    "learning_rate": 5e-05,
+    "num_train_epochs": 10
+  }
+}

run-1/checkpoint-14394/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4312a997194e916559618a59f3bbf4e1f98c127e89ae3b7fb5162b782863b12c
+size 5368

run-1/checkpoint-16793/config.json

@@ -0,0 +1,125 @@
+{
+  "architectures": [
+    "BertForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "classifier_dropout": null,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "LABEL_0",
+    "1": "LABEL_1",
+    "2": "LABEL_2",
+    "3": "LABEL_3",
+    "4": "LABEL_4",
+    "5": "LABEL_5",
+    "6": "LABEL_6",
+    "7": "LABEL_7",
+    "8": "LABEL_8",
+    "9": "LABEL_9",
+    "10": "LABEL_10",
+    "11": "LABEL_11",
+    "12": "LABEL_12",
+    "13": "LABEL_13",
+    "14": "LABEL_14",
+    "15": "LABEL_15",
+    "16": "LABEL_16",
+    "17": "LABEL_17",
+    "18": "LABEL_18",
+    "19": "LABEL_19",
+    "20": "LABEL_20",
+    "21": "LABEL_21",
+    "22": "LABEL_22",
+    "23": "LABEL_23",
+    "24": "LABEL_24",
+    "25": "LABEL_25",
+    "26": "LABEL_26",
+    "27": "LABEL_27",
+    "28": "LABEL_28",
+    "29": "LABEL_29",
+    "30": "LABEL_30",
+    "31": "LABEL_31",
+    "32": "LABEL_32",
+    "33": "LABEL_33",
+    "34": "LABEL_34",
+    "35": "LABEL_35",
+    "36": "LABEL_36",
+    "37": "LABEL_37",
+    "38": "LABEL_38",
+    "39": "LABEL_39",
+    "40": "LABEL_40",
+    "41": "LABEL_41",
+    "42": "LABEL_42",
+    "43": "LABEL_43",
+    "44": "LABEL_44",
+    "45": "LABEL_45",
+    "46": "LABEL_46",
+    "47": "LABEL_47"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "LABEL_0": 0,
+    "LABEL_1": 1,
+    "LABEL_10": 10,
+    "LABEL_11": 11,
+    "LABEL_12": 12,
+    "LABEL_13": 13,
+    "LABEL_14": 14,
+    "LABEL_15": 15,
+    "LABEL_16": 16,
+    "LABEL_17": 17,
+    "LABEL_18": 18,
+    "LABEL_19": 19,
+    "LABEL_2": 2,
+    "LABEL_20": 20,
+    "LABEL_21": 21,
+    "LABEL_22": 22,
+    "LABEL_23": 23,
+    "LABEL_24": 24,
+    "LABEL_25": 25,
+    "LABEL_26": 26,
+    "LABEL_27": 27,
+    "LABEL_28": 28,
+    "LABEL_29": 29,
+    "LABEL_3": 3,
+    "LABEL_30": 30,
+    "LABEL_31": 31,
+    "LABEL_32": 32,
+    "LABEL_33": 33,
+    "LABEL_34": 34,
+    "LABEL_35": 35,
+    "LABEL_36": 36,
+    "LABEL_37": 37,
+    "LABEL_38": 38,
+    "LABEL_39": 39,
+    "LABEL_4": 4,
+    "LABEL_40": 40,
+    "LABEL_41": 41,
+    "LABEL_42": 42,
+    "LABEL_43": 43,
+    "LABEL_44": 44,
+    "LABEL_45": 45,
+    "LABEL_46": 46,
+    "LABEL_47": 47,
+    "LABEL_5": 5,
+    "LABEL_6": 6,
+    "LABEL_7": 7,
+    "LABEL_8": 8,
+    "LABEL_9": 9
+  },
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "torch_dtype": "float32",
+  "transformers_version": "4.50.3",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 30522
+}

run-1/checkpoint-16793/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:c10b90ad6ab001474cc764f40a80231ed24ce00f70bb06826fede4251c93269a
+size 438100144

run-1/checkpoint-16793/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:e9c3ae1434af2354dc5eb533a65b574ce9aa0d35d9de9d0fbb91da18cdb6b1c4
+size 876321402

run-1/checkpoint-16793/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:428bd6bed852b4f8932455db9c461eeba79d3efe1fc880d20395baef552a0694
+size 14244

run-1/checkpoint-16793/scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f5dda3f875adbc098424e41c8f647dfbf085055fa6a9997f248634b057103d57
+size 988

run-1/checkpoint-16793/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:b46550d5556a9a8429669ce0755d27ffe05e7fe651b066b9e7b01d62f58ebe96
+size 1064

run-1/checkpoint-16793/trainer_state.json

@@ -0,0 +1,1454 @@
+{
+  "best_global_step": 11995,
+  "best_metric": 0.9374392372783462,
+  "best_model_checkpoint": "./results/run-1/checkpoint-11995",
+  "epoch": 7.0,
+  "eval_steps": 500,
+  "global_step": 16793,
+  "is_hyper_param_search": true,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.041684035014589414,
+      "grad_norm": 2.362386465072632,
+      "learning_rate": 9.900000000000002e-06,
+      "loss": 1.3944,
+      "step": 100
+    },
+    {
+      "epoch": 0.08336807002917883,
+      "grad_norm": 0.7483571767807007,
+      "learning_rate": 1.9900000000000003e-05,
+      "loss": 0.5576,
+      "step": 200
+    },
+    {
+      "epoch": 0.12505210504376824,
+      "grad_norm": 0.8563070297241211,
+      "learning_rate": 2.9900000000000002e-05,
+      "loss": 0.4502,
+      "step": 300
+    },
+    {
+      "epoch": 0.16673614005835766,
+      "grad_norm": 0.7200814485549927,
+      "learning_rate": 3.99e-05,
+      "loss": 0.388,
+      "step": 400
+    },
+    {
+      "epoch": 0.20842017507294705,
+      "grad_norm": 1.4553083181381226,
+      "learning_rate": 4.99e-05,
+      "loss": 0.3508,
+      "step": 500
+    },
+    {
+      "epoch": 0.25010421008753647,
+      "grad_norm": 0.9915399551391602,
+      "learning_rate": 4.978927203065135e-05,
+      "loss": 0.3186,
+      "step": 600
+    },
+    {
+      "epoch": 0.29178824510212586,
+      "grad_norm": 0.9978141188621521,
+      "learning_rate": 4.957641549595573e-05,
+      "loss": 0.3087,
+      "step": 700
+    },
+    {
+      "epoch": 0.3334722801167153,
+      "grad_norm": 1.3709461688995361,
+      "learning_rate": 4.936355896126011e-05,
+      "loss": 0.2976,
+      "step": 800
+    },
+    {
+      "epoch": 0.3751563151313047,
+      "grad_norm": 1.82745361328125,
+      "learning_rate": 4.91507024265645e-05,
+      "loss": 0.28,
+      "step": 900
+    },
+    {
+      "epoch": 0.4168403501458941,
+      "grad_norm": 1.2072925567626953,
+      "learning_rate": 4.893784589186888e-05,
+      "loss": 0.2751,
+      "step": 1000
+    },
+    {
+      "epoch": 0.45852438516048355,
+      "grad_norm": 1.0433937311172485,
+      "learning_rate": 4.872498935717327e-05,
+      "loss": 0.2712,
+      "step": 1100
+    },
+    {
+      "epoch": 0.5002084201750729,
+      "grad_norm": 1.1097549200057983,
+      "learning_rate": 4.851213282247765e-05,
+      "loss": 0.2909,
+      "step": 1200
+    },
+    {
+      "epoch": 0.5418924551896623,
+      "grad_norm": 1.2917771339416504,
+      "learning_rate": 4.8299276287782034e-05,
+      "loss": 0.2501,
+      "step": 1300
+    },
+    {
+      "epoch": 0.5835764902042517,
+      "grad_norm": 0.9812286496162415,
+      "learning_rate": 4.808641975308642e-05,
+      "loss": 0.2621,
+      "step": 1400
+    },
+    {
+      "epoch": 0.6252605252188412,
+      "grad_norm": 1.1786612272262573,
+      "learning_rate": 4.7873563218390804e-05,
+      "loss": 0.2435,
+      "step": 1500
+    },
+    {
+      "epoch": 0.6669445602334306,
+      "grad_norm": 1.4759145975112915,
+      "learning_rate": 4.766070668369519e-05,
+      "loss": 0.2398,
+      "step": 1600
+    },
+    {
+      "epoch": 0.70862859524802,
+      "grad_norm": 1.5548242330551147,
+      "learning_rate": 4.744785014899958e-05,
+      "loss": 0.2432,
+      "step": 1700
+    },
+    {
+      "epoch": 0.7503126302626094,
+      "grad_norm": 1.9005539417266846,
+      "learning_rate": 4.723499361430396e-05,
+      "loss": 0.2351,
+      "step": 1800
+    },
+    {
+      "epoch": 0.7919966652771988,
+      "grad_norm": 1.5831588506698608,
+      "learning_rate": 4.702213707960835e-05,
+      "loss": 0.2378,
+      "step": 1900
+    },
+    {
+      "epoch": 0.8336807002917882,
+      "grad_norm": 1.5482088327407837,
+      "learning_rate": 4.680928054491273e-05,
+      "loss": 0.2271,
+      "step": 2000
+    },
+    {
+      "epoch": 0.8753647353063777,
+      "grad_norm": 1.1437913179397583,
+      "learning_rate": 4.6596424010217114e-05,
+      "loss": 0.231,
+      "step": 2100
+    },
+    {
+      "epoch": 0.9170487703209671,
+      "grad_norm": 0.7985406517982483,
+      "learning_rate": 4.63835674755215e-05,
+      "loss": 0.228,
+      "step": 2200
+    },
+    {
+      "epoch": 0.9587328053355565,
+      "grad_norm": 1.9323142766952515,
+      "learning_rate": 4.6170710940825884e-05,
+      "loss": 0.2198,
+      "step": 2300
+    },
+    {
+      "epoch": 1.0,
+      "eval_administration_accuracy": 0.9322493224932249,
+      "eval_administration_f1": 0.9122911143651031,
+      "eval_corruption_accuracy": 0.9560141755263707,
+      "eval_corruption_f1": 0.9517304794081561,
+      "eval_democracy_accuracy": 0.9505941213258287,
+      "eval_democracy_f1": 0.9401838742047087,
+      "eval_development_accuracy": 0.9118198874296435,
+      "eval_development_f1": 0.9003553411809495,
+      "eval_economy_accuracy": 0.9264123410464874,
+      "eval_economy_f1": 0.9202465382357672,
+      "eval_education_accuracy": 0.9643527204502814,
+      "eval_education_f1": 0.9615509617901187,
+      "eval_environment_accuracy": 0.9799874921826142,
+      "eval_environment_f1": 0.9786142985651582,
+      "eval_instability_accuracy": 0.9417344173441734,
+      "eval_instability_f1": 0.9340921018628459,
+      "eval_leadership_accuracy": 0.8085261621846987,
+      "eval_leadership_f1": 0.7858417402549625,
+      "eval_loss": 0.21971678733825684,
+      "eval_overall_accuracy": 0.9324317281634354,
+      "eval_overall_f1": 0.9244250028140656,
+      "eval_race_accuracy": 0.9508025849489264,
+      "eval_race_f1": 0.9481714649378918,
+      "eval_religion_accuracy": 0.9530956848030019,
+      "eval_religion_f1": 0.9518106778632144,
+      "eval_runtime": 10.9084,
+      "eval_safety_accuracy": 0.9135918282259746,
+      "eval_safety_f1": 0.9082114410999095,
+      "eval_samples_per_second": 879.505,
+      "eval_steps_per_second": 55.003,
+      "step": 2399
+    },
+    {
+      "epoch": 1.0004168403501459,
+      "grad_norm": 1.2817295789718628,
+      "learning_rate": 4.5957854406130265e-05,
+      "loss": 0.2397,
+      "step": 2400
+    },
+    {
+      "epoch": 1.0421008753647354,
+      "grad_norm": 1.3623204231262207,
+      "learning_rate": 4.5744997871434654e-05,
+      "loss": 0.1885,
+      "step": 2500
+    },
+    {
+      "epoch": 1.0837849103793247,
+      "grad_norm": 1.2534806728363037,
+      "learning_rate": 4.553214133673904e-05,
+      "loss": 0.1843,
+      "step": 2600
+    },
+    {
+      "epoch": 1.1254689453939142,
+      "grad_norm": 1.2860488891601562,
+      "learning_rate": 4.5319284802043424e-05,
+      "loss": 0.1817,
+      "step": 2700
+    },
+    {
+      "epoch": 1.1671529804085035,
+      "grad_norm": 1.2887414693832397,
+      "learning_rate": 4.510642826734781e-05,
+      "loss": 0.1798,
+      "step": 2800
+    },
+    {
+      "epoch": 1.208837015423093,
+      "grad_norm": 1.973421335220337,
+      "learning_rate": 4.4893571732652194e-05,
+      "loss": 0.1878,
+      "step": 2900
+    },
+    {
+      "epoch": 1.2505210504376825,
+      "grad_norm": 0.7910940051078796,
+      "learning_rate": 4.468071519795658e-05,
+      "loss": 0.182,
+      "step": 3000
+    },
+    {
+      "epoch": 1.2922050854522718,
+      "grad_norm": 1.7422680854797363,
+      "learning_rate": 4.4467858663260964e-05,
+      "loss": 0.1833,
+      "step": 3100
+    },
+    {
+      "epoch": 1.3338891204668613,
+      "grad_norm": 1.1493451595306396,
+      "learning_rate": 4.4255002128565345e-05,
+      "loss": 0.1766,
+      "step": 3200
+    },
+    {
+      "epoch": 1.3755731554814505,
+      "grad_norm": 2.179731845855713,
+      "learning_rate": 4.4042145593869734e-05,
+      "loss": 0.1785,
+      "step": 3300
+    },
+    {
+      "epoch": 1.41725719049604,
+      "grad_norm": 1.5773117542266846,
+      "learning_rate": 4.3829289059174115e-05,
+      "loss": 0.1901,
+      "step": 3400
+    },
+    {
+      "epoch": 1.4589412255106295,
+      "grad_norm": 2.4681854248046875,
+      "learning_rate": 4.3616432524478504e-05,
+      "loss": 0.1833,
+      "step": 3500
+    },
+    {
+      "epoch": 1.5006252605252188,
+      "grad_norm": 2.391803503036499,
+      "learning_rate": 4.340357598978289e-05,
+      "loss": 0.1677,
+      "step": 3600
+    },
+    {
+      "epoch": 1.5423092955398081,
+      "grad_norm": 0.8182645440101624,
+      "learning_rate": 4.3190719455087274e-05,
+      "loss": 0.1666,
+      "step": 3700
+    },
+    {
+      "epoch": 1.5839933305543976,
+      "grad_norm": 1.2136952877044678,
+      "learning_rate": 4.297786292039166e-05,
+      "loss": 0.1787,
+      "step": 3800
+    },
+    {
+      "epoch": 1.6256773655689871,
+      "grad_norm": 2.435816764831543,
+      "learning_rate": 4.2765006385696044e-05,
+      "loss": 0.1678,
+      "step": 3900
+    },
+    {
+      "epoch": 1.6673614005835766,
+      "grad_norm": 1.5172932147979736,
+      "learning_rate": 4.2552149851000425e-05,
+      "loss": 0.1779,
+      "step": 4000
+    },
+    {
+      "epoch": 1.709045435598166,
+      "grad_norm": 1.4677119255065918,
+      "learning_rate": 4.2339293316304814e-05,
+      "loss": 0.1824,
+      "step": 4100
+    },
+    {
+      "epoch": 1.7507294706127552,
+      "grad_norm": 1.8327163457870483,
+      "learning_rate": 4.2126436781609195e-05,
+      "loss": 0.1719,
+      "step": 4200
+    },
+    {
+      "epoch": 1.7924135056273447,
+      "grad_norm": 1.2324726581573486,
+      "learning_rate": 4.1913580246913584e-05,
+      "loss": 0.1592,
+      "step": 4300
+    },
+    {
+      "epoch": 1.8340975406419342,
+      "grad_norm": 3.6276469230651855,
+      "learning_rate": 4.1700723712217965e-05,
+      "loss": 0.1636,
+      "step": 4400
+    },
+    {
+      "epoch": 1.8757815756565237,
+      "grad_norm": 0.940485417842865,
+      "learning_rate": 4.148786717752235e-05,
+      "loss": 0.1639,
+      "step": 4500
+    },
+    {
+      "epoch": 1.917465610671113,
+      "grad_norm": 2.002577781677246,
+      "learning_rate": 4.127501064282674e-05,
+      "loss": 0.1756,
+      "step": 4600
+    },
+    {
+      "epoch": 1.9591496456857023,
+      "grad_norm": 2.6062567234039307,
+      "learning_rate": 4.1062154108131124e-05,
+      "loss": 0.1729,
+      "step": 4700
+    },
+    {
+      "epoch": 2.0,
+      "eval_administration_accuracy": 0.9374609130706691,
+      "eval_administration_f1": 0.9302102302828306,
+      "eval_corruption_accuracy": 0.9603919116114238,
+      "eval_corruption_f1": 0.9585025895331236,
+      "eval_democracy_accuracy": 0.9562226391494684,
+      "eval_democracy_f1": 0.9505005443286667,
+      "eval_development_accuracy": 0.9199499687304565,
+      "eval_development_f1": 0.9145013995891959,
+      "eval_economy_accuracy": 0.9319366270585783,
+      "eval_economy_f1": 0.9302256375627648,
+      "eval_education_accuracy": 0.9651865749426725,
+      "eval_education_f1": 0.9641689686262792,
+      "eval_environment_accuracy": 0.9806128830519074,
+      "eval_environment_f1": 0.9798389855050841,
+      "eval_instability_accuracy": 0.9244319366270586,
+      "eval_instability_f1": 0.9301250908302109,
+      "eval_leadership_accuracy": 0.8395872420262664,
+      "eval_leadership_f1": 0.830858580138579,
+      "eval_loss": 0.20445191860198975,
+      "eval_overall_accuracy": 0.9366791744840525,
+      "eval_overall_f1": 0.9342598776015526,
+      "eval_race_accuracy": 0.9545549301646863,
+      "eval_race_f1": 0.9531323667533257,
+      "eval_religion_accuracy": 0.9575776526996039,
+      "eval_religion_f1": 0.9560680494194458,
+      "eval_runtime": 10.9745,
+      "eval_safety_accuracy": 0.9122368146758391,
+      "eval_safety_f1": 0.9129860886491253,
+      "eval_samples_per_second": 874.208,
+      "eval_steps_per_second": 54.672,
+      "step": 4798
+    },
+    {
+      "epoch": 2.0008336807002918,
+      "grad_norm": 1.0561970472335815,
+      "learning_rate": 4.085142613878246e-05,
+      "loss": 0.174,
+      "step": 4800
+    },
+    {
+      "epoch": 2.0425177157148813,
+      "grad_norm": 0.9414407014846802,
+      "learning_rate": 4.0638569604086845e-05,
+      "loss": 0.1259,
+      "step": 4900
+    },
+    {
+      "epoch": 2.084201750729471,
+      "grad_norm": 1.0318489074707031,
+      "learning_rate": 4.042571306939123e-05,
+      "loss": 0.1258,
+      "step": 5000
+    },
+    {
+      "epoch": 2.12588578574406,
+      "grad_norm": 2.4367470741271973,
+      "learning_rate": 4.0212856534695615e-05,
+      "loss": 0.1186,
+      "step": 5100
+    },
+    {
+      "epoch": 2.1675698207586493,
+      "grad_norm": 1.8129918575286865,
+      "learning_rate": 4e-05,
+      "loss": 0.1243,
+      "step": 5200
+    },
+    {
+      "epoch": 2.209253855773239,
+      "grad_norm": 2.235281467437744,
+      "learning_rate": 3.978714346530439e-05,
+      "loss": 0.1169,
+      "step": 5300
+    },
+    {
+      "epoch": 2.2509378907878284,
+      "grad_norm": 3.7830569744110107,
+      "learning_rate": 3.957428693060877e-05,
+      "loss": 0.1177,
+      "step": 5400
+    },
+    {
+      "epoch": 2.292621925802418,
+      "grad_norm": 1.2535426616668701,
+      "learning_rate": 3.9361430395913155e-05,
+      "loss": 0.1204,
+      "step": 5500
+    },
+    {
+      "epoch": 2.334305960817007,
+      "grad_norm": 1.443596601486206,
+      "learning_rate": 3.914857386121754e-05,
+      "loss": 0.1187,
+      "step": 5600
+    },
+    {
+      "epoch": 2.3759899958315964,
+      "grad_norm": 1.0241508483886719,
+      "learning_rate": 3.8935717326521925e-05,
+      "loss": 0.1222,
+      "step": 5700
+    },
+    {
+      "epoch": 2.417674030846186,
+      "grad_norm": 1.923004150390625,
+      "learning_rate": 3.8722860791826307e-05,
+      "loss": 0.1246,
+      "step": 5800
+    },
+    {
+      "epoch": 2.4593580658607754,
+      "grad_norm": 0.7686344385147095,
+      "learning_rate": 3.8510004257130695e-05,
+      "loss": 0.13,
+      "step": 5900
+    },
+    {
+      "epoch": 2.501042100875365,
+      "grad_norm": 1.762736439704895,
+      "learning_rate": 3.8297147722435077e-05,
+      "loss": 0.1263,
+      "step": 6000
+    },
+    {
+      "epoch": 2.542726135889954,
+      "grad_norm": 1.9006534814834595,
+      "learning_rate": 3.8084291187739465e-05,
+      "loss": 0.1152,
+      "step": 6100
+    },
+    {
+      "epoch": 2.5844101709045435,
+      "grad_norm": 2.003476858139038,
+      "learning_rate": 3.787143465304385e-05,
+      "loss": 0.1318,
+      "step": 6200
+    },
+    {
+      "epoch": 2.626094205919133,
+      "grad_norm": 1.8864182233810425,
+      "learning_rate": 3.7658578118348235e-05,
+      "loss": 0.1296,
+      "step": 6300
+    },
+    {
+      "epoch": 2.6677782409337225,
+      "grad_norm": 1.567968726158142,
+      "learning_rate": 3.744572158365262e-05,
+      "loss": 0.1292,
+      "step": 6400
+    },
+    {
+      "epoch": 2.7094622759483116,
+      "grad_norm": 1.7688707113265991,
+      "learning_rate": 3.7232865048957005e-05,
+      "loss": 0.1258,
+      "step": 6500
+    },
+    {
+      "epoch": 2.751146310962901,
+      "grad_norm": 2.138125419616699,
+      "learning_rate": 3.7020008514261387e-05,
+      "loss": 0.1208,
+      "step": 6600
+    },
+    {
+      "epoch": 2.7928303459774906,
+      "grad_norm": 3.3148305416107178,
+      "learning_rate": 3.6807151979565775e-05,
+      "loss": 0.1253,
+      "step": 6700
+    },
+    {
+      "epoch": 2.83451438099208,
+      "grad_norm": 1.8787152767181396,
+      "learning_rate": 3.6594295444870157e-05,
+      "loss": 0.1161,
+      "step": 6800
+    },
+    {
+      "epoch": 2.8761984160066696,
+      "grad_norm": 0.7995413541793823,
+      "learning_rate": 3.6381438910174545e-05,
+      "loss": 0.1196,
+      "step": 6900
+    },
+    {
+      "epoch": 2.917882451021259,
+      "grad_norm": 1.973639965057373,
+      "learning_rate": 3.6168582375478927e-05,
+      "loss": 0.1227,
+      "step": 7000
+    },
+    {
+      "epoch": 2.959566486035848,
+      "grad_norm": 1.0217853784561157,
+      "learning_rate": 3.5955725840783315e-05,
+      "loss": 0.125,
+      "step": 7100
+    },
+    {
+      "epoch": 3.0,
+      "eval_administration_accuracy": 0.9373566812591203,
+      "eval_administration_f1": 0.9333482432043958,
+      "eval_corruption_accuracy": 0.9609130706691682,
+      "eval_corruption_f1": 0.958848962871965,
+      "eval_democracy_accuracy": 0.9567437982072129,
+      "eval_democracy_f1": 0.9532298494465137,
+      "eval_development_accuracy": 0.9128622055451324,
+      "eval_development_f1": 0.9113351170841872,
+      "eval_economy_accuracy": 0.9317281634354805,
+      "eval_economy_f1": 0.9292396676492435,
+      "eval_education_accuracy": 0.965603502188868,
+      "eval_education_f1": 0.9637981071832747,
+      "eval_environment_accuracy": 0.9804044194288096,
+      "eval_environment_f1": 0.9788214035384287,
+      "eval_instability_accuracy": 0.9490306441525954,
+      "eval_instability_f1": 0.9469525951807101,
+      "eval_leadership_accuracy": 0.8446946007921617,
+      "eval_leadership_f1": 0.840338137042441,
+      "eval_loss": 0.20486579835414886,
+      "eval_overall_accuracy": 0.9398235007991106,
+      "eval_overall_f1": 0.9372835834491008,
+      "eval_race_accuracy": 0.9540337711069419,
+      "eval_race_f1": 0.9520143591935892,
+      "eval_religion_accuracy": 0.959349593495935,
+      "eval_religion_f1": 0.9575545048864775,
+      "eval_runtime": 10.9937,
+      "eval_safety_accuracy": 0.9251615593079008,
+      "eval_safety_f1": 0.9219220541079818,
+      "eval_samples_per_second": 872.682,
+      "eval_steps_per_second": 54.577,
+      "step": 7197
+    },
+    {
+      "epoch": 3.0012505210504377,
+      "grad_norm": 0.8939034938812256,
+      "learning_rate": 3.5744997871434655e-05,
+      "loss": 0.1224,
+      "step": 7200
+    },
+    {
+      "epoch": 3.042934556065027,
+      "grad_norm": 2.387557029724121,
+      "learning_rate": 3.5532141336739036e-05,
+      "loss": 0.0924,
+      "step": 7300
+    },
+    {
+      "epoch": 3.0846185910796167,
+      "grad_norm": 0.7587341666221619,
+      "learning_rate": 3.5319284802043425e-05,
+      "loss": 0.0838,
+      "step": 7400
+    },
+    {
+      "epoch": 3.1263026260942057,
+      "grad_norm": 2.06026291847229,
+      "learning_rate": 3.5106428267347806e-05,
+      "loss": 0.0894,
+      "step": 7500
+    },
+    {
+      "epoch": 3.1679866611087952,
+      "grad_norm": 2.001525402069092,
+      "learning_rate": 3.4893571732652195e-05,
+      "loss": 0.0913,
+      "step": 7600
+    },
+    {
+      "epoch": 3.2096706961233847,
+      "grad_norm": 2.0007410049438477,
+      "learning_rate": 3.468071519795658e-05,
+      "loss": 0.0891,
+      "step": 7700
+    },
+    {
+      "epoch": 3.2513547311379742,
+      "grad_norm": 1.6416945457458496,
+      "learning_rate": 3.4467858663260965e-05,
+      "loss": 0.0919,
+      "step": 7800
+    },
+    {
+      "epoch": 3.2930387661525637,
+      "grad_norm": 1.653064489364624,
+      "learning_rate": 3.425500212856535e-05,
+      "loss": 0.0932,
+      "step": 7900
+    },
+    {
+      "epoch": 3.334722801167153,
+      "grad_norm": 1.7365379333496094,
+      "learning_rate": 3.4042145593869735e-05,
+      "loss": 0.0898,
+      "step": 8000
+    },
+    {
+      "epoch": 3.3764068361817423,
+      "grad_norm": 1.561340093612671,
+      "learning_rate": 3.3829289059174116e-05,
+      "loss": 0.0866,
+      "step": 8100
+    },
+    {
+      "epoch": 3.418090871196332,
+      "grad_norm": 0.7200934886932373,
+      "learning_rate": 3.3616432524478505e-05,
+      "loss": 0.0844,
+      "step": 8200
+    },
+    {
+      "epoch": 3.4597749062109213,
+      "grad_norm": 0.8205280303955078,
+      "learning_rate": 3.3405704555129844e-05,
+      "loss": 0.0914,
+      "step": 8300
+    },
+    {
+      "epoch": 3.5014589412255104,
+      "grad_norm": 1.7642834186553955,
+      "learning_rate": 3.319284802043423e-05,
+      "loss": 0.0873,
+      "step": 8400
+    },
+    {
+      "epoch": 3.5431429762401,
+      "grad_norm": 2.3324527740478516,
+      "learning_rate": 3.2979991485738614e-05,
+      "loss": 0.0808,
+      "step": 8500
+    },
+    {
+      "epoch": 3.5848270112546894,
+      "grad_norm": 1.7013909816741943,
+      "learning_rate": 3.2767134951043e-05,
+      "loss": 0.094,
+      "step": 8600
+    },
+    {
+      "epoch": 3.626511046269279,
+      "grad_norm": 4.724839210510254,
+      "learning_rate": 3.2554278416347384e-05,
+      "loss": 0.0878,
+      "step": 8700
+    },
+    {
+      "epoch": 3.6681950812838684,
+      "grad_norm": 3.1913349628448486,
+      "learning_rate": 3.2341421881651766e-05,
+      "loss": 0.0918,
+      "step": 8800
+    },
+    {
+      "epoch": 3.709879116298458,
+      "grad_norm": 2.2082605361938477,
+      "learning_rate": 3.2128565346956154e-05,
+      "loss": 0.0846,
+      "step": 8900
+    },
+    {
+      "epoch": 3.751563151313047,
+      "grad_norm": 2.2428789138793945,
+      "learning_rate": 3.1915708812260536e-05,
+      "loss": 0.0878,
+      "step": 9000
+    },
+    {
+      "epoch": 3.7932471863276365,
+      "grad_norm": 2.0431671142578125,
+      "learning_rate": 3.170285227756492e-05,
+      "loss": 0.0964,
+      "step": 9100
+    },
+    {
+      "epoch": 3.834931221342226,
+      "grad_norm": 1.9629932641983032,
+      "learning_rate": 3.1489995742869306e-05,
+      "loss": 0.0969,
+      "step": 9200
+    },
+    {
+      "epoch": 3.8766152563568155,
+      "grad_norm": 1.0876961946487427,
+      "learning_rate": 3.1277139208173694e-05,
+      "loss": 0.0864,
+      "step": 9300
+    },
+    {
+      "epoch": 3.9182992913714045,
+      "grad_norm": 2.003713607788086,
+      "learning_rate": 3.1064282673478076e-05,
+      "loss": 0.0843,
+      "step": 9400
+    },
+    {
+      "epoch": 3.959983326385994,
+      "grad_norm": 1.3784735202789307,
+      "learning_rate": 3.0851426138782464e-05,
+      "loss": 0.0866,
+      "step": 9500
+    },
+    {
+      "epoch": 4.0,
+      "eval_administration_accuracy": 0.9397540129247446,
+      "eval_administration_f1": 0.9326250155599194,
+      "eval_corruption_accuracy": 0.9600792161767772,
+      "eval_corruption_f1": 0.9585031567143241,
+      "eval_democracy_accuracy": 0.9542422347300397,
+      "eval_democracy_f1": 0.9522572809405702,
+      "eval_development_accuracy": 0.9216176777152387,
+      "eval_development_f1": 0.9176939504856475,
+      "eval_economy_accuracy": 0.9336043360433605,
+      "eval_economy_f1": 0.9317409802418596,
+      "eval_education_accuracy": 0.968313529289139,
+      "eval_education_f1": 0.9659804103888047,
+      "eval_environment_accuracy": 0.9789451740671253,
+      "eval_environment_f1": 0.9783634505314646,
+      "eval_instability_accuracy": 0.9435063581405044,
+      "eval_instability_f1": 0.9430673845492272,
+      "eval_leadership_accuracy": 0.8415676464456953,
+      "eval_leadership_f1": 0.8394569407264164,
+      "eval_loss": 0.22299017012119293,
+      "eval_overall_accuracy": 0.9397540129247446,
+      "eval_overall_f1": 0.9374286167544477,
+      "eval_race_accuracy": 0.9551803210339795,
+      "eval_race_f1": 0.9543972425939777,
+      "eval_religion_accuracy": 0.9580988117573483,
+      "eval_religion_f1": 0.9574052840679498,
+      "eval_runtime": 10.9597,
+      "eval_safety_accuracy": 0.9221388367729831,
+      "eval_safety_f1": 0.9176523042532135,
+      "eval_samples_per_second": 875.387,
+      "eval_steps_per_second": 54.746,
+      "step": 9596
+    },
+    {
+      "epoch": 4.0016673614005835,
+      "grad_norm": 0.7530078887939453,
+      "learning_rate": 3.0638569604086846e-05,
+      "loss": 0.0852,
+      "step": 9600
+    },
+    {
+      "epoch": 4.043351396415173,
+      "grad_norm": 1.286962866783142,
+      "learning_rate": 3.042571306939123e-05,
+      "loss": 0.0617,
+      "step": 9700
+    },
+    {
+      "epoch": 4.0850354314297626,
+      "grad_norm": 1.2875369787216187,
+      "learning_rate": 3.0212856534695616e-05,
+      "loss": 0.0663,
+      "step": 9800
+    },
+    {
+      "epoch": 4.126719466444352,
+      "grad_norm": 2.4919862747192383,
+      "learning_rate": 3e-05,
+      "loss": 0.0609,
+      "step": 9900
+    },
+    {
+      "epoch": 4.168403501458942,
+      "grad_norm": 2.174107074737549,
+      "learning_rate": 2.9787143465304386e-05,
+      "loss": 0.0588,
+      "step": 10000
+    },
+    {
+      "epoch": 4.210087536473531,
+      "grad_norm": 1.7784463167190552,
+      "learning_rate": 2.9574286930608767e-05,
+      "loss": 0.0701,
+      "step": 10100
+    },
+    {
+      "epoch": 4.25177157148812,
+      "grad_norm": 1.0707236528396606,
+      "learning_rate": 2.936143039591316e-05,
+      "loss": 0.0738,
+      "step": 10200
+    },
+    {
+      "epoch": 4.293455606502709,
+      "grad_norm": 0.875899612903595,
+      "learning_rate": 2.9148573861217544e-05,
+      "loss": 0.0665,
+      "step": 10300
+    },
+    {
+      "epoch": 4.335139641517299,
+      "grad_norm": 2.555934190750122,
+      "learning_rate": 2.8935717326521926e-05,
+      "loss": 0.0639,
+      "step": 10400
+    },
+    {
+      "epoch": 4.376823676531888,
+      "grad_norm": 0.7635720372200012,
+      "learning_rate": 2.872286079182631e-05,
+      "loss": 0.0707,
+      "step": 10500
+    },
+    {
+      "epoch": 4.418507711546478,
+      "grad_norm": 1.4782235622406006,
+      "learning_rate": 2.8510004257130696e-05,
+      "loss": 0.0661,
+      "step": 10600
+    },
+    {
+      "epoch": 4.460191746561067,
+      "grad_norm": 2.679727077484131,
+      "learning_rate": 2.829714772243508e-05,
+      "loss": 0.0697,
+      "step": 10700
+    },
+    {
+      "epoch": 4.501875781575657,
+      "grad_norm": 0.8129181861877441,
+      "learning_rate": 2.8084291187739466e-05,
+      "loss": 0.0569,
+      "step": 10800
+    },
+    {
+      "epoch": 4.543559816590246,
+      "grad_norm": 1.4562526941299438,
+      "learning_rate": 2.7871434653043847e-05,
+      "loss": 0.0661,
+      "step": 10900
+    },
+    {
+      "epoch": 4.585243851604836,
+      "grad_norm": 1.9270412921905518,
+      "learning_rate": 2.7658578118348232e-05,
+      "loss": 0.057,
+      "step": 11000
+    },
+    {
+      "epoch": 4.626927886619425,
+      "grad_norm": 1.7193882465362549,
+      "learning_rate": 2.7445721583652617e-05,
+      "loss": 0.069,
+      "step": 11100
+    },
+    {
+      "epoch": 4.668611921634014,
+      "grad_norm": 2.909707546234131,
+      "learning_rate": 2.7232865048957006e-05,
+      "loss": 0.0639,
+      "step": 11200
+    },
+    {
+      "epoch": 4.710295956648603,
+      "grad_norm": 0.570263683795929,
+      "learning_rate": 2.702000851426139e-05,
+      "loss": 0.0654,
+      "step": 11300
+    },
+    {
+      "epoch": 4.751979991663193,
+      "grad_norm": 0.43538907170295715,
+      "learning_rate": 2.6807151979565776e-05,
+      "loss": 0.0633,
+      "step": 11400
+    },
+    {
+      "epoch": 4.793664026677782,
+      "grad_norm": 2.182251214981079,
+      "learning_rate": 2.659429544487016e-05,
+      "loss": 0.0697,
+      "step": 11500
+    },
+    {
+      "epoch": 4.835348061692372,
+      "grad_norm": 2.25455904006958,
+      "learning_rate": 2.6381438910174542e-05,
+      "loss": 0.0599,
+      "step": 11600
+    },
+    {
+      "epoch": 4.877032096706961,
+      "grad_norm": 1.227161169052124,
+      "learning_rate": 2.6168582375478927e-05,
+      "loss": 0.0637,
+      "step": 11700
+    },
+    {
+      "epoch": 4.918716131721551,
+      "grad_norm": 1.0838350057601929,
+      "learning_rate": 2.5955725840783312e-05,
+      "loss": 0.0635,
+      "step": 11800
+    },
+    {
+      "epoch": 4.96040016673614,
+      "grad_norm": 1.7880126237869263,
+      "learning_rate": 2.5742869306087697e-05,
+      "loss": 0.0573,
+      "step": 11900
+    },
+    {
+      "epoch": 5.0,
+      "eval_administration_accuracy": 0.9370439858244737,
+      "eval_administration_f1": 0.9329617589185005,
+      "eval_corruption_accuracy": 0.9603919116114238,
+      "eval_corruption_f1": 0.9582673518936824,
+      "eval_democracy_accuracy": 0.9550760892224307,
+      "eval_democracy_f1": 0.9548460506714369,
+      "eval_development_accuracy": 0.9215134459036898,
+      "eval_development_f1": 0.9164255722085631,
+      "eval_economy_accuracy": 0.931311236189285,
+      "eval_economy_f1": 0.9304486638126732,
+      "eval_education_accuracy": 0.9673754429851991,
+      "eval_education_f1": 0.9662633072587713,
+      "eval_environment_accuracy": 0.9785282468209298,
+      "eval_environment_f1": 0.9784438171410429,
+      "eval_instability_accuracy": 0.9468417761100688,
+      "eval_instability_f1": 0.9453606763183978,
+      "eval_leadership_accuracy": 0.8380237648530331,
+      "eval_leadership_f1": 0.8360259652707104,
+      "eval_loss": 0.2365979105234146,
+      "eval_overall_accuracy": 0.9393631436314364,
+      "eval_overall_f1": 0.9374392372783462,
+      "eval_race_accuracy": 0.9540337711069419,
+      "eval_race_f1": 0.9532943942985637,
+      "eval_religion_accuracy": 0.960287679799875,
+      "eval_religion_f1": 0.9582749968466362,
+      "eval_runtime": 10.946,
+      "eval_safety_accuracy": 0.9219303731498854,
+      "eval_safety_f1": 0.9186582927011756,
+      "eval_samples_per_second": 876.487,
+      "eval_steps_per_second": 54.815,
+      "step": 11995
+    },
+    {
+      "epoch": 5.00208420175073,
+      "grad_norm": 1.8319880962371826,
+      "learning_rate": 2.5530012771392082e-05,
+      "loss": 0.0644,
+      "step": 12000
+    },
+    {
+      "epoch": 5.0437682367653185,
+      "grad_norm": 0.9142705798149109,
+      "learning_rate": 2.5317156236696464e-05,
+      "loss": 0.0468,
+      "step": 12100
+    },
+    {
+      "epoch": 5.085452271779908,
+      "grad_norm": 1.5496271848678589,
+      "learning_rate": 2.5104299702000856e-05,
+      "loss": 0.0513,
+      "step": 12200
+    },
+    {
+      "epoch": 5.1271363067944975,
+      "grad_norm": 1.026939034461975,
+      "learning_rate": 2.4891443167305237e-05,
+      "loss": 0.0505,
+      "step": 12300
+    },
+    {
+      "epoch": 5.168820341809087,
+      "grad_norm": 0.9045308232307434,
+      "learning_rate": 2.4678586632609622e-05,
+      "loss": 0.0423,
+      "step": 12400
+    },
+    {
+      "epoch": 5.2105043768236765,
+      "grad_norm": 1.3580307960510254,
+      "learning_rate": 2.4465730097914007e-05,
+      "loss": 0.044,
+      "step": 12500
+    },
+    {
+      "epoch": 5.252188411838266,
+      "grad_norm": 2.3576512336730957,
+      "learning_rate": 2.4252873563218392e-05,
+      "loss": 0.0495,
+      "step": 12600
+    },
+    {
+      "epoch": 5.2938724468528555,
+      "grad_norm": 0.727289080619812,
+      "learning_rate": 2.4040017028522777e-05,
+      "loss": 0.0501,
+      "step": 12700
+    },
+    {
+      "epoch": 5.335556481867445,
+      "grad_norm": 0.7748624086380005,
+      "learning_rate": 2.3827160493827162e-05,
+      "loss": 0.0479,
+      "step": 12800
+    },
+    {
+      "epoch": 5.3772405168820345,
+      "grad_norm": 1.302977204322815,
+      "learning_rate": 2.3614303959131544e-05,
+      "loss": 0.046,
+      "step": 12900
+    },
+    {
+      "epoch": 5.418924551896623,
+      "grad_norm": 1.363124132156372,
+      "learning_rate": 2.3401447424435932e-05,
+      "loss": 0.0432,
+      "step": 13000
+    },
+    {
+      "epoch": 5.460608586911213,
+      "grad_norm": 0.7924700379371643,
+      "learning_rate": 2.3188590889740317e-05,
+      "loss": 0.0458,
+      "step": 13100
+    },
+    {
+      "epoch": 5.502292621925802,
+      "grad_norm": 0.6111019849777222,
+      "learning_rate": 2.29757343550447e-05,
+      "loss": 0.0466,
+      "step": 13200
+    },
+    {
+      "epoch": 5.543976656940392,
+      "grad_norm": 1.055407166481018,
+      "learning_rate": 2.2762877820349084e-05,
+      "loss": 0.0487,
+      "step": 13300
+    },
+    {
+      "epoch": 5.585660691954981,
+      "grad_norm": 0.47858452796936035,
+      "learning_rate": 2.255002128565347e-05,
+      "loss": 0.0448,
+      "step": 13400
+    },
+    {
+      "epoch": 5.627344726969571,
+      "grad_norm": 2.17866587638855,
+      "learning_rate": 2.2339293316304812e-05,
+      "loss": 0.0454,
+      "step": 13500
+    },
+    {
+      "epoch": 5.66902876198416,
+      "grad_norm": 0.9695106744766235,
+      "learning_rate": 2.2126436781609197e-05,
+      "loss": 0.0487,
+      "step": 13600
+    },
+    {
+      "epoch": 5.71071279699875,
+      "grad_norm": 1.223860502243042,
+      "learning_rate": 2.1913580246913582e-05,
+      "loss": 0.0464,
+      "step": 13700
+    },
+    {
+      "epoch": 5.752396832013339,
+      "grad_norm": 0.8663516640663147,
+      "learning_rate": 2.1700723712217967e-05,
+      "loss": 0.044,
+      "step": 13800
+    },
+    {
+      "epoch": 5.794080867027929,
+      "grad_norm": 0.7240914106369019,
+      "learning_rate": 2.148786717752235e-05,
+      "loss": 0.0463,
+      "step": 13900
+    },
+    {
+      "epoch": 5.835764902042518,
+      "grad_norm": 1.0169471502304077,
+      "learning_rate": 2.1275010642826737e-05,
+      "loss": 0.0449,
+      "step": 14000
+    },
+    {
+      "epoch": 5.877448937057107,
+      "grad_norm": 2.1470980644226074,
+      "learning_rate": 2.1062154108131122e-05,
+      "loss": 0.0387,
+      "step": 14100
+    },
+    {
+      "epoch": 5.919132972071696,
+      "grad_norm": 0.8170746564865112,
+      "learning_rate": 2.0849297573435507e-05,
+      "loss": 0.0444,
+      "step": 14200
+    },
+    {
+      "epoch": 5.960817007086286,
+      "grad_norm": 1.886773943901062,
+      "learning_rate": 2.063644103873989e-05,
+      "loss": 0.0408,
+      "step": 14300
+    },
+    {
+      "epoch": 6.0,
+      "eval_administration_accuracy": 0.9392328538670002,
+      "eval_administration_f1": 0.9335586868984106,
+      "eval_corruption_accuracy": 0.9608088388576194,
+      "eval_corruption_f1": 0.9586222965202317,
+      "eval_democracy_accuracy": 0.9575776526996039,
+      "eval_democracy_f1": 0.9552908906267366,
+      "eval_development_accuracy": 0.9196372732958099,
+      "eval_development_f1": 0.9170852123784591,
+      "eval_economy_accuracy": 0.9321450906816761,
+      "eval_economy_f1": 0.9289418561764905,
+      "eval_education_accuracy": 0.9687304565353346,
+      "eval_education_f1": 0.9666551400031382,
+      "eval_environment_accuracy": 0.9798832603710652,
+      "eval_environment_f1": 0.9786999163300606,
+      "eval_instability_accuracy": 0.9454867625599332,
+      "eval_instability_f1": 0.9422890312692406,
+      "eval_leadership_accuracy": 0.8316656243485512,
+      "eval_leadership_f1": 0.833450633159851,
+      "eval_loss": 0.2633351981639862,
+      "eval_overall_accuracy": 0.9391546800083385,
+      "eval_overall_f1": 0.937151862792397,
+      "eval_race_accuracy": 0.9554930164686263,
+      "eval_race_f1": 0.9546556151565829,
+      "eval_religion_accuracy": 0.9575776526996039,
+      "eval_religion_f1": 0.9567690967660134,
+      "eval_runtime": 11.0311,
+      "eval_safety_accuracy": 0.9216176777152387,
+      "eval_safety_f1": 0.9198039782235481,
+      "eval_samples_per_second": 869.723,
+      "eval_steps_per_second": 54.392,
+      "step": 14394
+    },
+    {
+      "epoch": 6.002501042100875,
+      "grad_norm": 0.6738269329071045,
+      "learning_rate": 2.0423584504044274e-05,
+      "loss": 0.0483,
+      "step": 14400
+    },
+    {
+      "epoch": 6.044185077115465,
+      "grad_norm": 1.050459623336792,
+      "learning_rate": 2.0210727969348662e-05,
+      "loss": 0.0343,
+      "step": 14500
+    },
+    {
+      "epoch": 6.085869112130054,
+      "grad_norm": 0.7071917653083801,
+      "learning_rate": 1.9997871434653047e-05,
+      "loss": 0.0303,
+      "step": 14600
+    },
+    {
+      "epoch": 6.127553147144644,
+      "grad_norm": 1.2754780054092407,
+      "learning_rate": 1.978501489995743e-05,
+      "loss": 0.0347,
+      "step": 14700
+    },
+    {
+      "epoch": 6.169237182159233,
+      "grad_norm": 1.0299218893051147,
+      "learning_rate": 1.9572158365261814e-05,
+      "loss": 0.0319,
+      "step": 14800
+    },
+    {
+      "epoch": 6.210921217173823,
+      "grad_norm": 1.029572606086731,
+      "learning_rate": 1.93593018305662e-05,
+      "loss": 0.0322,
+      "step": 14900
+    },
+    {
+      "epoch": 6.2526052521884115,
+      "grad_norm": 2.746795654296875,
+      "learning_rate": 1.9146445295870584e-05,
+      "loss": 0.0348,
+      "step": 15000
+    },
+    {
+      "epoch": 6.294289287203001,
+      "grad_norm": 0.7789388298988342,
+      "learning_rate": 1.893358876117497e-05,
+      "loss": 0.035,
+      "step": 15100
+    },
+    {
+      "epoch": 6.3359733222175905,
+      "grad_norm": 0.9597777724266052,
+      "learning_rate": 1.8720732226479354e-05,
+      "loss": 0.0352,
+      "step": 15200
+    },
+    {
+      "epoch": 6.37765735723218,
+      "grad_norm": 0.322014719247818,
+      "learning_rate": 1.850787569178374e-05,
+      "loss": 0.0293,
+      "step": 15300
+    },
+    {
+      "epoch": 6.4193413922467695,
+      "grad_norm": 0.8756703734397888,
+      "learning_rate": 1.8295019157088124e-05,
+      "loss": 0.0253,
+      "step": 15400
+    },
+    {
+      "epoch": 6.461025427261359,
+      "grad_norm": 1.5358089208602905,
+      "learning_rate": 1.808216262239251e-05,
+      "loss": 0.0333,
+      "step": 15500
+    },
+    {
+      "epoch": 6.5027094622759485,
+      "grad_norm": 0.836966872215271,
+      "learning_rate": 1.7869306087696894e-05,
+      "loss": 0.0317,
+      "step": 15600
+    },
+    {
+      "epoch": 6.544393497290538,
+      "grad_norm": 0.5500324964523315,
+      "learning_rate": 1.765644955300128e-05,
+      "loss": 0.0354,
+      "step": 15700
+    },
+    {
+      "epoch": 6.5860775323051275,
+      "grad_norm": 0.5586037635803223,
+      "learning_rate": 1.7443593018305664e-05,
+      "loss": 0.0365,
+      "step": 15800
+    },
+    {
+      "epoch": 6.627761567319716,
+      "grad_norm": 0.8741154074668884,
+      "learning_rate": 1.7230736483610045e-05,
+      "loss": 0.0354,
+      "step": 15900
+    },
+    {
+      "epoch": 6.669445602334306,
+      "grad_norm": 0.2844066023826599,
+      "learning_rate": 1.7017879948914434e-05,
+      "loss": 0.0342,
+      "step": 16000
+    },
+    {
+      "epoch": 6.711129637348895,
+      "grad_norm": 0.37321197986602783,
+      "learning_rate": 1.680502341421882e-05,
+      "loss": 0.03,
+      "step": 16100
+    },
+    {
+      "epoch": 6.752813672363485,
+      "grad_norm": 1.4438360929489136,
+      "learning_rate": 1.6592166879523204e-05,
+      "loss": 0.0339,
+      "step": 16200
+    },
+    {
+      "epoch": 6.794497707378074,
+      "grad_norm": 0.8285302519798279,
+      "learning_rate": 1.6379310344827585e-05,
+      "loss": 0.0281,
+      "step": 16300
+    },
+    {
+      "epoch": 6.836181742392664,
+      "grad_norm": 2.7729620933532715,
+      "learning_rate": 1.616645381013197e-05,
+      "loss": 0.0317,
+      "step": 16400
+    },
+    {
+      "epoch": 6.877865777407253,
+      "grad_norm": 0.46065399050712585,
+      "learning_rate": 1.595359727543636e-05,
+      "loss": 0.0316,
+      "step": 16500
+    },
+    {
+      "epoch": 6.919549812421843,
+      "grad_norm": 0.254247784614563,
+      "learning_rate": 1.574074074074074e-05,
+      "loss": 0.031,
+      "step": 16600
+    },
+    {
+      "epoch": 6.961233847436432,
+      "grad_norm": 1.5957903861999512,
+      "learning_rate": 1.5527884206045125e-05,
+      "loss": 0.0346,
+      "step": 16700
+    },
+    {
+      "epoch": 7.0,
+      "eval_administration_accuracy": 0.9354805086512403,
+      "eval_administration_f1": 0.9319059881529553,
+      "eval_corruption_accuracy": 0.9599749843652282,
+      "eval_corruption_f1": 0.9578990365929126,
+      "eval_democracy_accuracy": 0.9550760892224307,
+      "eval_democracy_f1": 0.9543016568544691,
+      "eval_development_accuracy": 0.9124452782989368,
+      "eval_development_f1": 0.9099602216486293,
+      "eval_economy_accuracy": 0.9293308317698562,
+      "eval_economy_f1": 0.9281251063602204,
+      "eval_education_accuracy": 0.9676881384198457,
+      "eval_education_f1": 0.9660833921396531,
+      "eval_environment_accuracy": 0.9796747967479674,
+      "eval_environment_f1": 0.9786146756936395,
+      "eval_instability_accuracy": 0.9451740671252866,
+      "eval_instability_f1": 0.9429347319872009,
+      "eval_leadership_accuracy": 0.8364602876797999,
+      "eval_leadership_f1": 0.8378023994018406,
+      "eval_loss": 0.27679628133773804,
+      "eval_overall_accuracy": 0.937460913070669,
+      "eval_overall_f1": 0.9361318238368415,
+      "eval_race_accuracy": 0.9520533666875131,
+      "eval_race_f1": 0.952198677089417,
+      "eval_religion_accuracy": 0.9548676255993329,
+      "eval_religion_f1": 0.9552183154626219,
+      "eval_runtime": 11.019,
+      "eval_safety_accuracy": 0.921304982280592,
+      "eval_safety_f1": 0.9185376846585376,
+      "eval_samples_per_second": 870.675,
+      "eval_steps_per_second": 54.451,
+      "step": 16793
+    }
+  ],
+  "logging_steps": 100,
+  "max_steps": 23990,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 10,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 2,
+        "early_stopping_threshold": 0.0
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 2
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": true
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 7.031335850960486e+16,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": {
+    "gradient_accumulation_steps": 1,
+    "learning_rate": 5e-05,
+    "num_train_epochs": 10
+  }
+}

run-1/checkpoint-16793/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:4312a997194e916559618a59f3bbf4e1f98c127e89ae3b7fb5162b782863b12c
+size 5368

run-2/checkpoint-1200/config.json

@@ -0,0 +1,125 @@
+{
+  "architectures": [
+    "BertForSequenceClassification"
+  ],
+  "attention_probs_dropout_prob": 0.1,
+  "classifier_dropout": null,
+  "gradient_checkpointing": false,
+  "hidden_act": "gelu",
+  "hidden_dropout_prob": 0.1,
+  "hidden_size": 768,
+  "id2label": {
+    "0": "LABEL_0",
+    "1": "LABEL_1",
+    "2": "LABEL_2",
+    "3": "LABEL_3",
+    "4": "LABEL_4",
+    "5": "LABEL_5",
+    "6": "LABEL_6",
+    "7": "LABEL_7",
+    "8": "LABEL_8",
+    "9": "LABEL_9",
+    "10": "LABEL_10",
+    "11": "LABEL_11",
+    "12": "LABEL_12",
+    "13": "LABEL_13",
+    "14": "LABEL_14",
+    "15": "LABEL_15",
+    "16": "LABEL_16",
+    "17": "LABEL_17",
+    "18": "LABEL_18",
+    "19": "LABEL_19",
+    "20": "LABEL_20",
+    "21": "LABEL_21",
+    "22": "LABEL_22",
+    "23": "LABEL_23",
+    "24": "LABEL_24",
+    "25": "LABEL_25",
+    "26": "LABEL_26",
+    "27": "LABEL_27",
+    "28": "LABEL_28",
+    "29": "LABEL_29",
+    "30": "LABEL_30",
+    "31": "LABEL_31",
+    "32": "LABEL_32",
+    "33": "LABEL_33",
+    "34": "LABEL_34",
+    "35": "LABEL_35",
+    "36": "LABEL_36",
+    "37": "LABEL_37",
+    "38": "LABEL_38",
+    "39": "LABEL_39",
+    "40": "LABEL_40",
+    "41": "LABEL_41",
+    "42": "LABEL_42",
+    "43": "LABEL_43",
+    "44": "LABEL_44",
+    "45": "LABEL_45",
+    "46": "LABEL_46",
+    "47": "LABEL_47"
+  },
+  "initializer_range": 0.02,
+  "intermediate_size": 3072,
+  "label2id": {
+    "LABEL_0": 0,
+    "LABEL_1": 1,
+    "LABEL_10": 10,
+    "LABEL_11": 11,
+    "LABEL_12": 12,
+    "LABEL_13": 13,
+    "LABEL_14": 14,
+    "LABEL_15": 15,
+    "LABEL_16": 16,
+    "LABEL_17": 17,
+    "LABEL_18": 18,
+    "LABEL_19": 19,
+    "LABEL_2": 2,
+    "LABEL_20": 20,
+    "LABEL_21": 21,
+    "LABEL_22": 22,
+    "LABEL_23": 23,
+    "LABEL_24": 24,
+    "LABEL_25": 25,
+    "LABEL_26": 26,
+    "LABEL_27": 27,
+    "LABEL_28": 28,
+    "LABEL_29": 29,
+    "LABEL_3": 3,
+    "LABEL_30": 30,
+    "LABEL_31": 31,
+    "LABEL_32": 32,
+    "LABEL_33": 33,
+    "LABEL_34": 34,
+    "LABEL_35": 35,
+    "LABEL_36": 36,
+    "LABEL_37": 37,
+    "LABEL_38": 38,
+    "LABEL_39": 39,
+    "LABEL_4": 4,
+    "LABEL_40": 40,
+    "LABEL_41": 41,
+    "LABEL_42": 42,
+    "LABEL_43": 43,
+    "LABEL_44": 44,
+    "LABEL_45": 45,
+    "LABEL_46": 46,
+    "LABEL_47": 47,
+    "LABEL_5": 5,
+    "LABEL_6": 6,
+    "LABEL_7": 7,
+    "LABEL_8": 8,
+    "LABEL_9": 9
+  },
+  "layer_norm_eps": 1e-12,
+  "max_position_embeddings": 512,
+  "model_type": "bert",
+  "num_attention_heads": 12,
+  "num_hidden_layers": 12,
+  "pad_token_id": 0,
+  "position_embedding_type": "absolute",
+  "torch_dtype": "float32",
+  "transformers_version": "4.50.3",
+  "type_vocab_size": 2,
+  "use_cache": true,
+  "vocab_size": 30522
+}

run-2/checkpoint-1200/model.safetensors

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d9a9413e0558b2b03f34ffaeef8d03315ff63af1209bdfcd55acf4ce04908097
+size 438100144

run-2/checkpoint-1200/optimizer.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:270fb7ab21afb9ed6b14d5c4477c432803b6a66033e6f3a2c9f1abb1a16fe095
+size 876321402

run-2/checkpoint-1200/rng_state.pth

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:03f6521411fa3fb31926db64cfd06b37e9d7c0607f2eec0bf790c73df4910854
+size 14244

run-2/checkpoint-1200/scaler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:942cfd7aded7d16363d1ae1a2911c01ef4e25f3c70ed059c88f1845d9b6c24dc
+size 988

run-2/checkpoint-1200/scheduler.pt

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:50ff1bd41ce0d0f4d18181c5a3e67a33e2df006cf69c336e1cb8a88cc1d930f6
+size 1064

run-2/checkpoint-1200/trainer_state.json

@@ -0,0 +1,165 @@
+{
+  "best_global_step": 1200,
+  "best_metric": 0.9199542168750479,
+  "best_model_checkpoint": "./results/run-2/checkpoint-1200",
+  "epoch": 1.0,
+  "eval_steps": 500,
+  "global_step": 1200,
+  "is_hyper_param_search": true,
+  "is_local_process_zero": true,
+  "is_world_process_zero": true,
+  "log_history": [
+    {
+      "epoch": 0.08336807002917883,
+      "grad_norm": 1.2878578901290894,
+      "learning_rate": 1.4e-05,
+      "loss": 1.2644,
+      "step": 100
+    },
+    {
+      "epoch": 0.16673614005835766,
+      "grad_norm": 1.0009047985076904,
+      "learning_rate": 2.8e-05,
+      "loss": 0.4758,
+      "step": 200
+    },
+    {
+      "epoch": 0.25010421008753647,
+      "grad_norm": 0.7635347843170166,
+      "learning_rate": 4.2e-05,
+      "loss": 0.4054,
+      "step": 300
+    },
+    {
+      "epoch": 0.3334722801167153,
+      "grad_norm": 0.8168604373931885,
+      "learning_rate": 5.6e-05,
+      "loss": 0.3468,
+      "step": 400
+    },
+    {
+      "epoch": 0.4168403501458941,
+      "grad_norm": 1.1069319248199463,
+      "learning_rate": 7e-05,
+      "loss": 0.3051,
+      "step": 500
+    },
+    {
+      "epoch": 0.5002084201750729,
+      "grad_norm": 1.015086054801941,
+      "learning_rate": 6.962534789124383e-05,
+      "loss": 0.3044,
+      "step": 600
+    },
+    {
+      "epoch": 0.5835764902042517,
+      "grad_norm": 0.8430781364440918,
+      "learning_rate": 6.925069578248769e-05,
+      "loss": 0.2742,
+      "step": 700
+    },
+    {
+      "epoch": 0.6669445602334306,
+      "grad_norm": 0.950409471988678,
+      "learning_rate": 6.887604367373152e-05,
+      "loss": 0.2573,
+      "step": 800
+    },
+    {
+      "epoch": 0.7503126302626094,
+      "grad_norm": 1.171885371208191,
+      "learning_rate": 6.850139156497538e-05,
+      "loss": 0.2517,
+      "step": 900
+    },
+    {
+      "epoch": 0.8336807002917882,
+      "grad_norm": 0.9133914709091187,
+      "learning_rate": 6.812673945621922e-05,
+      "loss": 0.2435,
+      "step": 1000
+    },
+    {
+      "epoch": 0.9170487703209671,
+      "grad_norm": 0.7939682602882385,
+      "learning_rate": 6.775208734746307e-05,
+      "loss": 0.2373,
+      "step": 1100
+    },
+    {
+      "epoch": 1.0,
+      "grad_norm": 0.49670639634132385,
+      "learning_rate": 6.737743523870691e-05,
+      "loss": 0.2363,
+      "step": 1200
+    },
+    {
+      "epoch": 1.0,
+      "eval_administration_accuracy": 0.9302689180737961,
+      "eval_administration_f1": 0.9077189587684091,
+      "eval_corruption_accuracy": 0.9534083802376485,
+      "eval_corruption_f1": 0.9474431988592255,
+      "eval_democracy_accuracy": 0.9473629351678132,
+      "eval_democracy_f1": 0.9358042521135516,
+      "eval_development_accuracy": 0.9094225557640192,
+      "eval_development_f1": 0.8937523747915405,
+      "eval_economy_accuracy": 0.924119241192412,
+      "eval_economy_f1": 0.9156575066932242,
+      "eval_education_accuracy": 0.9621638524077548,
+      "eval_education_f1": 0.9583809787976314,
+      "eval_environment_accuracy": 0.9775901605169898,
+      "eval_environment_f1": 0.9754739686963207,
+      "eval_instability_accuracy": 0.9422555764019178,
+      "eval_instability_f1": 0.9343996801103537,
+      "eval_leadership_accuracy": 0.7970606629143214,
+      "eval_leadership_f1": 0.7702375495381067,
+      "eval_loss": 0.22745274007320404,
+      "eval_overall_accuracy": 0.9296261552359112,
+      "eval_overall_f1": 0.9199542168750479,
+      "eval_race_accuracy": 0.9505941213258287,
+      "eval_race_f1": 0.9483610849847235,
+      "eval_religion_accuracy": 0.9505941213258287,
+      "eval_religion_f1": 0.9489643730683611,
+      "eval_runtime": 11.0331,
+      "eval_safety_accuracy": 0.9106733375026058,
+      "eval_safety_f1": 0.9032566760791263,
+      "eval_samples_per_second": 869.566,
+      "eval_steps_per_second": 54.382,
+      "step": 1200
+    }
+  ],
+  "logging_steps": 100,
+  "max_steps": 19184,
+  "num_input_tokens_seen": 0,
+  "num_train_epochs": 16,
+  "save_steps": 500,
+  "stateful_callbacks": {
+    "EarlyStoppingCallback": {
+      "args": {
+        "early_stopping_patience": 2,
+        "early_stopping_threshold": 0.0
+      },
+      "attributes": {
+        "early_stopping_patience_counter": 0
+      }
+    },
+    "TrainerControl": {
+      "args": {
+        "should_epoch_stop": false,
+        "should_evaluate": false,
+        "should_log": false,
+        "should_save": true,
+        "should_training_stop": false
+      },
+      "attributes": {}
+    }
+  },
+  "total_flos": 1.0100267260379136e+16,
+  "train_batch_size": 16,
+  "trial_name": null,
+  "trial_params": {
+    "gradient_accumulation_steps": 2,
+    "learning_rate": 7e-05,
+    "num_train_epochs": 16
+  }
+}

run-2/checkpoint-1200/training_args.bin

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:a417688ba3a38960888d219af94bafd1d9551df9005996513f2ac72dfc76c145
+size 5368

training_args.bin

@@ -1,3 +1,3 @@
 version https://git-lfs.github.com/spec/v1
-oid sha256:4312a997194e916559618a59f3bbf4e1f98c127e89ae3b7fb5162b782863b12c
+oid sha256:a417688ba3a38960888d219af94bafd1d9551df9005996513f2ac72dfc76c145
 size 5368