artifacts/patch_back/results/tiiuae__falcon-7b-instruct/layer24/_summary/summary.csv

kind,file,task,eval_mode,layer,seed,hf_id,hf_split,candidate_labels,Qs_shape,patch_desc,donor_source,donor_tasks,donor_pick,n_donor_bank,scan_effective,scan_skipped,base_acc_scan,ablt_acc_scan,flips_scan,anti_flips_scan,both_correct_scan,both_wrong_scan,patched_primary_method,patched_primary_rescued_pct,patched_primary_mean_dmargin,diff_time_shuffled_minus_patched_primary_rescued_pct,diff_patched_primary_minus_shared_randvec_rescued_pct,patched_0_rescued,patched_0_n,patched_0_rescued_pct,patched_0_mean_dmargin,patched_01_rescued,patched_01_n,patched_01_rescued_pct,patched_01_mean_dmargin,patched_full_rescued,patched_full_n,patched_full_rescued_pct,patched_full_mean_dmargin,patched_self_rescued,patched_self_n,patched_self_rescued_pct,patched_self_mean_dmargin,patched_transfer_rescued,patched_transfer_n,patched_transfer_rescued_pct,patched_transfer_mean_dmargin,control_time_shuffled_rescued,control_time_shuffled_n,control_time_shuffled_rescued_pct,control_time_shuffled_mean_dmargin,control_shared_mismatch_rescued,control_shared_mismatch_n,control_shared_mismatch_rescued_pct,control_shared_mismatch_mean_dmargin,control_shared_perm_rescued,control_shared_perm_n,control_shared_perm_rescued_pct,control_shared_perm_mean_dmargin,control_shared_signflip_rescued,control_shared_signflip_n,control_shared_signflip_rescued_pct,control_shared_signflip_mean_dmargin,control_shared_randvec_rescued,control_shared_randvec_n,control_shared_randvec_rescued_pct,control_shared_randvec_mean_dmargin,control_rand_subspace_rescued,control_rand_subspace_n,control_rand_subspace_rescued_pct,control_rand_subspace_mean_dmargin,control_patch_nonshared_rescued,control_patch_nonshared_n,control_patch_nonshared_rescued_pct,control_patch_nonshared_mean_dmargin
flipset,aqua_alpha_sweep_seed123.json,aqua,,24,123,,,ABCDE,4544x113,steps=0,,,,,254,0,0.25196850393700787,0.2125984251968504,44,34,20,156,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
flipset,aqua_alpha_sweep_seed456.json,aqua,,24,456,,,ABCDE,4544x133,steps=0,,,,,254,0,0.25196850393700787,0.21653543307086615,21,12,43,178,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,,
flipset,aqua_transfer_cross_mc_baselinecorrect_seed123.json,aqua,,24,123,,,ABCDE,4544x113,steps=0,cross_task_eval,"commonsenseqa,openbookqa",random,114,254,0,0.25196850393700787,0.2125984251968504,44,34,20,156,patched_self,100.0,2.120680093765259,,,,,,,,,,,,,,,44,44,100.0,2.120680093765259,40,44,90.9090909090909,2.1255545616149902,,,,,,,,,,,,,,,,,,,,,,,,,,,,
flipset,aqua_transfer_same_task_seed123.json,aqua,,24,123,,,ABCDE,4544x113,steps=0,same_task_eval,aqua,random,254,254,0,0.25196850393700787,0.2125984251968504,44,34,20,156,patched_self,100.0,2.120680093765259,,,,,,,,,,,,,,,44,44,100.0,2.120680093765259,41,44,93.18181818181819,2.083189010620117,,,,,,,,,,,,,,,,,,,,,,,,,,,,
openanswer,gsm8k_genmath.json,gsm8k,gen_math,24,123,gsm8k/main,,,4544x113,"steps=0,1,2,3",,,,,256,0,0.03515625,0.0234375,7,4,2,243,patched_self,14.285714285714286,,14.285714285714286,14.285714285714286,,,,,,,,,,,,,1,7,14.285714285714286,,,,,,2,7,28.571428571428573,,,,,,,,,,,,,,0,7,0.0,,0,7,0.0,,0,7,0.0,
openanswer,gsm8k_pairlogprob.json,gsm8k,pair_logprob,24,123,gsm8k/main,,,4544x113,"steps=0,1,2,3",,,,,256,0,0.63671875,0.6484375,17,20,146,73,patched_self,100.0,1.4387097358703613,-29.411764705882348,100.0,,,,,,,,,,,,,17,17,100.0,1.4387097358703613,,,,,12,17,70.58823529411765,1.372696042060852,,,,,,,,,,,,,0,17,0.0,-0.04551048204302788,0,17,0.0,-0.0028862743638455868,0,17,0.0,-0.010198768228292465
openanswer,humaneval_gencode_compile.json,humaneval,gen_code_compile,24,123,openai_humaneval,test,,4544x113,"steps=0,1,2,3",,,,,0,164,,,0,0,0,0,patched_self,20.0,,-1.8181818181818166,9.090909090909092,,,,,,,,,,,,,11,55,20.0,,,,,,10,55,18.181818181818183,,,,,,,,,,,,,,6,55,10.909090909090908,,0,55,0.0,,1,55,1.8181818181818181,
openanswer,humaneval_pairlogprob.json,humaneval,pair_logprob,24,123,openai_humaneval,test,,4544x113,"steps=0,1,2,3",,,,,164,0,0.6402439024390244,0.6402439024390244,1,1,104,58,patched_self,100.0,0.6100447177886963,0.0,100.0,,,,,,,,,,,,,1,1,100.0,0.6100447177886963,,,,,1,1,100.0,0.6100447177886963,,,,,,,,,,,,,0,1,0.0,0.020630836486816406,0,1,0.0,-0.0474238395690918,0,1,0.0,1.1920928955078125e-06
subspace_mc,aqua.json,aqua,,24,123,,,ABCDE,4544x113,,,,,,254,0,0.25196850393700787,0.2125984251968504,44,34,20,156,patched_0,100.0,2.120680107311769,-4.545454545454547,84.0909090909091,44,44,100.0,2.120680107311769,44,44,100.0,2.120680107311769,44,44,100.0,2.120680107311769,,,,,,,,,42,44,95.45454545454545,2.110366859219291,,,,,,,,,,,,,7,44,15.909090909090908,0.3961740948937156,4,44,9.090909090909092,-0.17473465204238892,0,44,0.0,7.044185291637074e-08
subspace_mc,arc_challenge.json,arc_challenge,,24,123,,,ABCD,4544x113,,,,,,256,0,0.26171875,0.2109375,32,19,35,170,patched_0,100.0,1.6439488716423512,-28.125,90.625,32,32,100.0,1.6439488716423512,32,32,100.0,1.6439488716423512,32,32,100.0,1.6439488716423512,,,,,,,,,23,32,71.875,1.4819882325828075,,,,,,,,,,,,,3,32,9.375,-0.4834681339561939,2,32,6.25,-0.5636857226490974,0,32,0.0,1.7136335372924805e-07
subspace_mc,commonsenseqa.json,commonsenseqa,,24,123,,,ABCDE,4544x113,,,,,,256,0,0.1875,0.2265625,29,39,19,169,patched_0,100.0,1.672135235934422,-10.34482758620689,82.75862068965517,29,29,100.0,1.672135235934422,29,29,100.0,1.672135235934422,29,29,100.0,1.672135235934422,,,,,,,,,26,29,89.65517241379311,1.6617727033023177,,,,,,,,,,,,,5,29,17.24137931034483,0.1979779128370614,5,29,17.24137931034483,-0.45122285546927615,0,29,0.0,-4.9327981883081896e-08
subspace_mc,logiqa.json,logiqa,,24,123,,,ABCD,4544x113,,,,,,256,0,0.16796875,0.265625,7,32,36,181,patched_0,100.0,1.802259922027588,-42.857142857142854,71.42857142857143,7,7,100.0,1.802259922027588,7,7,100.0,1.802259922027588,7,7,100.0,1.802259922027588,,,,,,,,,4,7,57.142857142857146,1.7422590936933244,,,,,,,,,,,,,2,7,28.571428571428573,0.31407383510044645,2,7,28.571428571428573,0.2836837427956717,0,7,0.0,-5.108969552176339e-07
subspace_mc,openbookqa.json,openbookqa,,24,123,,,ABCD,4544x113,,,,,,256,0,0.24609375,0.265625,25,30,38,163,patched_0,100.0,1.4053189063072205,-16.0,68.0,25,25,100.0,1.4053189063072205,25,25,100.0,1.4053189063072205,25,25,100.0,1.4053189063072205,,,,,,,,,21,25,84.0,1.2916813731193542,,,,,,,,,,,,,8,25,32.0,0.025307378768920898,6,25,24.0,-0.36917322158813476,0,25,0.0,3.147125244140625e-07
subspace_mc,piqa.json,piqa,,24,123,,,AB,4544x113,,,,,,256,0,0.5390625,0.57421875,38,47,100,71,patched_0,100.0,1.0669233140192533,-15.78947368421052,78.94736842105263,38,38,100.0,1.0669233140192533,38,38,100.0,1.0669233140192533,38,38,100.0,1.0669233140192533,,,,,,,,,32,38,84.21052631578948,1.041327859226026,,,,,,,,,,,,,8,38,21.05263157894737,-0.277350924516979,17,38,44.73684210526316,0.5642389937451011,0,38,0.0,-1.1920928955078125e-07
subspace_mc,qasc.json,qasc,,24,123,,,ABCDEFGH,4544x113,,,,,,256,0,0.12109375,0.12109375,17,17,14,208,patched_0,100.0,2.0982201169518864,-17.647058823529406,76.47058823529412,17,17,100.0,2.0982201169518864,17,17,100.0,2.0982201169518864,17,17,100.0,2.0982201169518864,,,,,,,,,14,17,82.3529411764706,1.974048894994399,,,,,,,,,,,,,4,17,23.529411764705884,0.48005182602826285,2,17,11.764705882352942,-0.08192226466010599,0,17,0.0,4.4878791360294117e-07