Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

MCQA

Benchmarks

Task NameDataset NameSOTA ResultTrend
Multiple Choice Question AnsweringMCQA
Accuracy93.5
25
Circuit LocalizationMCQA
CPR2.52
21
Circuit RecoveryMCQA
CPR (Faithfulness Score)1.13
14
Causal LocalizationMCQA (val)
Accuracy92
10
Counterfactual ReasoningMCQA (test)
AP (Exact Counterfactual Accuracy)90.75
6
Multiple-choice Question AnsweringMCQA events
Coverage80.56
2
Multiple-choice Question AnsweringMCQA landmarks
Coverage79.54
2
Multiple-choice Question AnsweringMCQA movies
Coverage81.79
2
Multiple-choice Question AnsweringMCQA cities
Coverage79.08
2
Multiple-choice Question AnsweringMCQA books
Coverage82.02
2
Multiple-choice Question AnsweringMCQA artworks
Coverage79.29
2
Multiple-choice Question AnsweringMCQA persons
Coverage81.37
2
Multiple-choice Question AnsweringMCQA inventions
Coverage84.59
2
Showing 13 of 13 rows