Add files using upload-large-folder tool

revision_runs/heldout_v3/trainonly/runs/base_4/bce_lr0.0004/model_roberta/dataset_20newsgroups/head_seed_1/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_bce_only,100.0,0.0,0.0,cls_head,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.9194367921640648,0.7308815717697144,0.8487703917190589,0.8651348484828117

revision_runs/heldout_v3/trainonly/runs/base_4/bce_lr0.0004/model_roberta/dataset_20newsgroups/head_seed_2/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_bce_only,100.0,0.0,0.0,cls_head,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.919895928986838,0.7308815717697144,0.8487703917190589,0.8680119925062518

revision_runs/heldout_v3/trainonly/runs/base_4/bce_lr0.0004/model_roberta/dataset_20newsgroups/head_seed_3/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_bce_only,100.0,0.0,0.0,cls_head,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.9148454239363331,0.7308815717697144,0.8487703917190589,0.8673848548680014

revision_runs/heldout_v3/trainonly/runs/base_4/bce_lr0.0004/model_roberta/dataset_20newsgroups/head_seed_4/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_bce_only,100.0,0.0,0.0,cls_head,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.917141108050199,0.7308815717697144,0.8487703917190589,0.8650567923406403

revision_runs/heldout_v3/trainonly/runs/base_4/confidnet_lr0.0003_ep30_wd0.0/model_roberta/dataset_20newsgroups/head_seed_1/LoraConfig_20newsgroups_seed_4_normalization_True_confidnet.csv

@@ -0,0 +1,2 @@
+model,adapter,dataset,train_on_dataset,normalization,seed,test-acc,valid-confidnet-roc-auc,confidnet_roc_auc,best_lr,best_epoch,best_weighting,best_hidden,best_weight_decay
+roberta-base,LoraConfig,20newsgroups,train,True,4,0.7308815717697144,0.9221916131007041,0.8255099750103845,0.0003,30,1.0,400,0.0

revision_runs/heldout_v3/trainonly/runs/base_4/confidnet_lr0.0003_ep30_wd0.0/model_roberta/dataset_20newsgroups/head_seed_2/LoraConfig_20newsgroups_seed_4_normalization_True_confidnet.csv

@@ -0,0 +1,2 @@
+model,adapter,dataset,train_on_dataset,normalization,seed,test-acc,valid-confidnet-roc-auc,confidnet_roc_auc,best_lr,best_epoch,best_weighting,best_hidden,best_weight_decay
+roberta-base,LoraConfig,20newsgroups,train,True,4,0.7308815717697144,0.9159779614325069,0.818532687913889,0.0003,30,1.0,400,0.0

revision_runs/heldout_v3/trainonly/runs/base_4/confidnet_lr0.0003_ep30_wd0.0/model_roberta/dataset_20newsgroups/head_seed_3/LoraConfig_20newsgroups_seed_4_normalization_True_confidnet.csv

@@ -0,0 +1,2 @@
+model,adapter,dataset,train_on_dataset,normalization,seed,test-acc,valid-confidnet-roc-auc,confidnet_roc_auc,best_lr,best_epoch,best_weighting,best_hidden,best_weight_decay
+roberta-base,LoraConfig,20newsgroups,train,True,4,0.7308815717697144,0.918732782369146,0.828115983720231,0.0003,30,1.0,400,0.0

revision_runs/heldout_v3/trainonly/runs/base_4/confidnet_lr0.0003_ep30_wd0.0/model_roberta/dataset_20newsgroups/head_seed_4/LoraConfig_20newsgroups_seed_4_normalization_True_confidnet.csv

@@ -0,0 +1,2 @@
+model,adapter,dataset,train_on_dataset,normalization,seed,test-acc,valid-confidnet-roc-auc,confidnet_roc_auc,best_lr,best_epoch,best_weighting,best_hidden,best_weight_decay
+roberta-base,LoraConfig,20newsgroups,train,True,4,0.7308815717697144,0.9201101928374655,0.8287853308222735,0.0003,30,1.0,400,0.0

revision_runs/heldout_v3/trainonly/runs/base_4/l2sp_lr0.0004_b0.01/model_roberta/dataset_20newsgroups/head_seed_1/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,entropy_response_finetune_l2sp_only,100.0,0.0,0.01,cls_head,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.9018365472910927,0.7308815717697144,0.8487703917190589,0.8579939212995137

revision_runs/heldout_v3/trainonly/runs/base_4/l2sp_lr0.0004_b0.01/model_roberta/dataset_20newsgroups/head_seed_2/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,entropy_response_finetune_l2sp_only,100.0,0.0,0.01,cls_head,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.8998163452708907,0.7308815717697144,0.8487703917190589,0.8605698636078696

revision_runs/heldout_v3/trainonly/runs/base_4/l2sp_lr0.0004_b0.01/model_roberta/dataset_20newsgroups/head_seed_3/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,entropy_response_finetune_l2sp_only,100.0,0.0,0.01,cls_head,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.9015916743189472,0.7308815717697144,0.8487703917190589,0.8617303102037122

revision_runs/heldout_v3/trainonly/runs/base_4/l2sp_lr0.0004_b0.01/model_roberta/dataset_20newsgroups/head_seed_4/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,entropy_response_finetune_l2sp_only,100.0,0.0,0.01,cls_head,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.9048056320783594,0.7308815717697144,0.8487703917190589,0.8601664988594035

revision_runs/heldout_v3/trainonly/runs/base_4/loo_nol2sp/model_electra/dataset_20newsgroups/head_seed_0/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,google/electra-base-discriminator,LoraConfig,20newsgroups,train,4,entropy_response_finetune,100.0,1.0,0.0,cls_head,0.0001,entropy,0,0.7748344540596008,0.8662085916987877,0.9024635495223731,0.7238449454307556,0.8389883387888708,0.8590766796941136

revision_runs/heldout_v3/trainonly/runs/base_4/loo_nol2sp/model_electra/dataset_20newsgroups/head_seed_1/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,google/electra-base-discriminator,LoraConfig,20newsgroups,train,4,entropy_response_finetune,100.0,1.0,0.0,cls_head,0.0001,entropy,0,0.7748344540596008,0.8662085916987877,0.9021842355175689,0.7238449454307556,0.8389883387888708,0.8589173344996895

revision_runs/heldout_v3/trainonly/runs/base_4/loo_nol2sp/model_electra/dataset_20newsgroups/head_seed_2/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,google/electra-base-discriminator,LoraConfig,20newsgroups,train,4,entropy_response_finetune,100.0,1.0,0.0,cls_head,0.0001,entropy,0,0.7748344540596008,0.8662085916987877,0.9008714596949892,0.7238449454307556,0.8389883387888708,0.8590536641458321

revision_runs/heldout_v3/trainonly/runs/base_4/loo_nol2sp/model_electra/dataset_20newsgroups/head_seed_3/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,google/electra-base-discriminator,LoraConfig,20newsgroups,train,4,entropy_response_finetune,100.0,1.0,0.0,cls_head,0.0001,entropy,0,0.7748344540596008,0.8662085916987877,0.9023518239204514,0.7238449454307556,0.8389883387888708,0.859636460155765

revision_runs/heldout_v3/trainonly/runs/base_4/loo_nol2sp/model_electra/dataset_20newsgroups/head_seed_4/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,google/electra-base-discriminator,LoraConfig,20newsgroups,train,4,entropy_response_finetune,100.0,1.0,0.0,cls_head,0.0001,entropy,0,0.7748344540596008,0.8662085916987877,0.9019049215127646,0.7238449454307556,0.8389883387888708,0.859192286528585

revision_runs/heldout_v3/trainonly/runs/base_4/loo_nol2sp/model_roberta/dataset_20newsgroups/head_seed_2/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,entropy_response_finetune,100.0,1.0,0.0,cls_head,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.9179675543311907,0.7308815717697144,0.8487703917190589,0.8687495379139114

revision_runs/heldout_v3/trainonly/runs/base_4/loo_nol2sp/model_roberta/dataset_20newsgroups/head_seed_3/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,entropy_response_finetune,100.0,1.0,0.0,cls_head,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.918732782369146,0.7308815717697144,0.8487703917190589,0.8713274517895782

revision_runs/heldout_v3/trainonly/runs/base_4/loo_nol2sp/model_roberta/dataset_20newsgroups/head_seed_4/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,entropy_response_finetune,100.0,1.0,0.0,cls_head,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.9182430364248546,0.7308815717697144,0.8487703917190589,0.8665064320143099

revision_runs/heldout_v3/trainonly/runs/base_4/loo_noreg/model_electra/dataset_20newsgroups/head_seed_0/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,google/electra-base-discriminator,LoraConfig,20newsgroups,train,4,entropy_response_finetune,100.0,0.0,0.01,cls_head,0.0001,entropy,0,0.7748344540596008,0.8662085916987877,0.8813753421596559,0.7238449454307556,0.8389883387888708,0.844472564760991

revision_runs/heldout_v3/trainonly/runs/base_4/loo_noreg/model_electra/dataset_20newsgroups/head_seed_1/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,google/electra-base-discriminator,LoraConfig,20newsgroups,train,4,entropy_response_finetune,100.0,0.0,0.01,cls_head,0.0001,entropy,0,0.7748344540596008,0.8662085916987877,0.8807608513490867,0.7238449454307556,0.8389883387888708,0.8447336721880467

revision_runs/heldout_v3/trainonly/runs/base_4/loo_noreg/model_electra/dataset_20newsgroups/head_seed_2/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,google/electra-base-discriminator,LoraConfig,20newsgroups,train,4,entropy_response_finetune,100.0,0.0,0.01,cls_head,0.0001,entropy,0,0.7748344540596008,0.8662085916987877,0.8844198648120216,0.7238449454307556,0.8389883387888708,0.8484263008634798

revision_runs/heldout_v3/trainonly/runs/base_4/loo_noreg/model_electra/dataset_20newsgroups/head_seed_3/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,google/electra-base-discriminator,LoraConfig,20newsgroups,train,4,entropy_response_finetune,100.0,0.0,0.01,cls_head,0.0001,entropy,0,0.7748344540596008,0.8662085916987877,0.8833026087928049,0.7238449454307556,0.8389883387888708,0.8502932057396015

revision_runs/heldout_v3/trainonly/runs/base_4/loo_noreg/model_electra/dataset_20newsgroups/head_seed_4/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,google/electra-base-discriminator,LoraConfig,20newsgroups,train,4,entropy_response_finetune,100.0,0.0,0.01,cls_head,0.0001,entropy,0,0.7748344540596008,0.8662085916987877,0.8821574213731076,0.7238449454307556,0.8389883387888708,0.8476090284440431

revision_runs/heldout_v3/trainonly/runs/base_4/loo_randcls/model_electra/dataset_20newsgroups/head_seed_0/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,google/electra-base-discriminator,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_cls_random_reg,100.0,1.0,0.01,cls_random,0.0001,entropy,0,0.7748344540596008,0.8662085916987877,0.8142003240042456,0.7238449454307556,0.8389883387888708,0.7743950702635588

revision_runs/heldout_v3/trainonly/runs/base_4/loo_randcls/model_electra/dataset_20newsgroups/head_seed_1/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,google/electra-base-discriminator,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_cls_random_reg,100.0,1.0,0.01,cls_random,0.0001,entropy,0,0.7748344540596008,0.8662085916987877,0.814032735601363,0.7238449454307556,0.8389883387888708,0.7711221887521869

revision_runs/heldout_v3/trainonly/runs/base_4/loo_randcls/model_electra/dataset_20newsgroups/head_seed_2/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,google/electra-base-discriminator,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_cls_random_reg,100.0,1.0,0.01,cls_random,0.0001,entropy,0,0.7748344540596008,0.8662085916987877,0.8131947935869506,0.7238449454307556,0.8389883387888708,0.7708402703312828

revision_runs/heldout_v3/trainonly/runs/base_4/loo_randcls/model_electra/dataset_20newsgroups/head_seed_3/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,google/electra-base-discriminator,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_cls_random_reg,100.0,1.0,0.01,cls_random,0.0001,entropy,0,0.7748344540596008,0.8662085916987877,0.811239595553321,0.7238449454307556,0.8389883387888708,0.7731923976381286

revision_runs/heldout_v3/trainonly/runs/base_4/loo_randcls/model_roberta/dataset_20newsgroups/head_seed_0/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_cls_random_reg,100.0,1.0,0.01,cls_random,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.8001224364860728,0.7308815717697144,0.8487703917190589,0.7330360747528708

revision_runs/heldout_v3/trainonly/runs/base_4/loo_randcls/model_roberta/dataset_20newsgroups/head_seed_1/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_cls_random_reg,100.0,1.0,0.01,cls_random,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.8031527395163758,0.7308815717697144,0.8487703917190589,0.7432410863873582

revision_runs/heldout_v3/trainonly/runs/base_4/loo_randcls/model_roberta/dataset_20newsgroups/head_seed_2/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_cls_random_reg,100.0,1.0,0.01,cls_random,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.8082338536883991,0.7308815717697144,0.8487703917190589,0.7408286049324133

revision_runs/heldout_v3/trainonly/runs/base_4/loo_randcls/model_roberta/dataset_20newsgroups/head_seed_3/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_cls_random_reg,100.0,1.0,0.01,cls_random,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.8146311600857055,0.7308815717697144,0.8487703917190589,0.7535820913579484

revision_runs/heldout_v3/trainonly/runs/base_4/loo_randcls/model_roberta/dataset_20newsgroups/head_seed_4/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_cls_random_reg,100.0,1.0,0.01,cls_random,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.8219161310070402,0.7308815717697144,0.8487703917190589,0.7445855160599841

revision_runs/heldout_v3/trainonly/runs/base_4/loo_randlin/model_roberta/dataset_20newsgroups/head_seed_0/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_linear_random_reg,100.0,1.0,0.01,linear_random,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.8074074074074074,0.7308815717697144,0.8487703917190589,0.7517928940233282

revision_runs/heldout_v3/trainonly/runs/base_4/loo_randlin/model_roberta/dataset_20newsgroups/head_seed_1/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_linear_random_reg,100.0,1.0,0.01,linear_random,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.8064279155188245,0.7308815717697144,0.8487703917190589,0.7497495885473446

revision_runs/heldout_v3/trainonly/runs/base_4/loo_randlin/model_roberta/dataset_20newsgroups/head_seed_2/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_linear_random_reg,100.0,1.0,0.01,linear_random,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.8096418732782369,0.7308815717697144,0.8487703917190589,0.752331535174329

revision_runs/heldout_v3/trainonly/runs/base_4/loo_randlin/model_roberta/dataset_20newsgroups/head_seed_4/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_linear_random_reg,100.0,1.0,0.01,linear_random,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.8004897459442915,0.7308815717697144,0.8487703917190589,0.75342145343046

revision_runs/heldout_v3/trainonly/runs/base_4/randcls_lr0.0004/model_roberta/dataset_20newsgroups/head_seed_1/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_bce_only_cls_random,100.0,0.0,0.0,cls_random,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.8549433731251913,0.7308815717697144,0.8487703917190589,0.781886494181412

revision_runs/heldout_v3/trainonly/runs/base_4/randcls_lr0.0004/model_roberta/dataset_20newsgroups/head_seed_2/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_bce_only_cls_random,100.0,0.0,0.0,cls_random,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.8619528619528619,0.7308815717697144,0.8487703917190589,0.7867800586720508

revision_runs/heldout_v3/trainonly/runs/base_4/randcls_lr0.0004/model_roberta/dataset_20newsgroups/head_seed_3/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_bce_only_cls_random,100.0,0.0,0.0,cls_random,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.8650137741046832,0.7308815717697144,0.8487703917190589,0.7868184594262648

revision_runs/heldout_v3/trainonly/runs/base_4/randcls_lr0.0004/model_roberta/dataset_20newsgroups/head_seed_4/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_bce_only_cls_random,100.0,0.0,0.0,cls_random,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.8688399142944597,0.7308815717697144,0.8487703917190589,0.7832409609239839

revision_runs/heldout_v3/trainonly/runs/base_4/randlin_lr0.0004/model_roberta/dataset_20newsgroups/head_seed_1/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_bce_only_linear_random,100.0,0.0,0.0,linear_random,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.8073155800428528,0.7308815717697144,0.8487703917190589,0.7503920506406025

revision_runs/heldout_v3/trainonly/runs/base_4/randlin_lr0.0004/model_roberta/dataset_20newsgroups/head_seed_2/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_bce_only_linear_random,100.0,0.0,0.0,linear_random,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.8101010101010101,0.7308815717697144,0.8487703917190589,0.7525053019477741

revision_runs/heldout_v3/trainonly/runs/base_4/randlin_lr0.0004/model_roberta/dataset_20newsgroups/head_seed_3/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_bce_only_linear_random,100.0,0.0,0.0,linear_random,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.808937863483318,0.7308815717697144,0.8487703917190589,0.7530482895085286

revision_runs/heldout_v3/trainonly/runs/base_4/randlin_lr0.0004/model_roberta/dataset_20newsgroups/head_seed_4/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,ablation_entropy_finetune_bce_only_linear_random,100.0,0.0,0.0,linear_random,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.8031221303948577,0.7308815717697144,0.8487703917190589,0.7531074813362029

revision_runs/heldout_v3/trainonly/runs/base_4/reg_lr0.0004_a0.1/model_roberta/dataset_20newsgroups/head_seed_1/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,entropy_response_finetune_reg_only,100.0,0.1,0.0,cls_head,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.9192837465564738,0.7308815717697144,0.8487703917190589,0.8655266974858484

revision_runs/heldout_v3/trainonly/runs/base_4/reg_lr0.0004_a0.1/model_roberta/dataset_20newsgroups/head_seed_2/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,entropy_response_finetune_reg_only,100.0,0.1,0.0,cls_head,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.9196816651362105,0.7308815717697144,0.8487703917190589,0.8681430121157292

revision_runs/heldout_v3/trainonly/runs/base_4/reg_lr0.0004_a0.1/model_roberta/dataset_20newsgroups/head_seed_3/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,entropy_response_finetune_reg_only,100.0,0.1,0.0,cls_head,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.9167737985919804,0.7308815717697144,0.8487703917190589,0.8681942728658121

revision_runs/heldout_v3/trainonly/runs/base_4/reg_lr0.0004_a0.1/model_roberta/dataset_20newsgroups/head_seed_4/best_run_metric_df.csv

@@ -0,0 +1,2 @@
+,model,adapter,dataset,train_on_dataset,seed,grid,lam,reg_alpha,l2sp_alpha,load_weights,lr,head_type,trial_number,valid-acc,valid-base-max-prob-roc-auc,valid-fine-tune-max-prob-roc-auc,test-acc,test-base-max-prob-roc-auc,test-fine-tune-max-prob-roc-auc
+0,roberta-base,LoraConfig,20newsgroups,train,4,entropy_response_finetune_reg_only,100.0,0.1,0.0,cls_head,0.0004,entropy,0,0.8013244867324829,0.8837159473523111,0.9176920722375268,0.7308815717697144,0.8487703917190589,0.8651872742499419