Task,Accuracy,Centered
hellaswag_zeroshot,0.433579,0.244772
jeopardy,0.027870,0.027870
bigbench_qa_wikidata,0.381477,0.381477
arc_easy,0.587963,0.450617
arc_challenge,0.308020,0.077361
copa,0.560000,0.120000
commonsense_qa,0.313677,0.142097
piqa,0.691513,0.383025
openbook_qa,0.360000,0.146667
lambada_openai,0.323501,0.323501
hellaswag,0.432185,0.242913
winograd,0.593407,0.186813
winogrande,0.522494,0.044988
bigbench_dyck_languages,0.199000,0.199000
agi_eval_lsat_ar,0.256522,0.070652
bigbench_cs_algorithms,0.431818,0.431818
bigbench_operators,0.152381,0.152381
bigbench_repeat_copy_logic,0.031250,0.031250
squad,0.097919,0.097919
coqa,0.141927,0.141927
boolq,0.537920,-0.215999
bigbench_language_identification,0.249200,0.174037
CORE,,0.175231