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

Common sense reasoning benchmarks

Benchmarks

Task NameDataset NameSOTA ResultTrend
Common-sense ReasoningCommon-sense Reasoning Benchmarks Zero-shot
BoolQ Accuracy80.1
16
Common Sense ReasoningSix common sense reasoning benchmarks (ARC-e, PIQA, OpenbookQA, Winogrande, HellaSwag, MathQA)
Average Accuracy61
15
Common-sense ReasoningCommon-sense Reasoning Benchmarks (BoolQ, SciQ, PIQA, WinoG., ARC-C, HellaS.) zero-shot
BoolQ Accuracy81
2
Showing 3 of 3 rows