ARC-Easy, ARC-Challenge, OpenBookQA, WinoGrande, PIQA, HellaSwag, MathQA, RTE, BoolQ
Benchmarks
Task NameDataset NameSOTA ResultTrendResults
ARC-Easy, ARC-Challenge, OpenBookQA, WinoGrande, PIQA, HellaSwag, MathQA, RTE, BoolQ zero-shot
71.08Mean Accuracy
59