ResearchDatasetsBoolQ, ARC, WinoGrande, HellaSwagBoolQ, ARC, WinoGrande, HellaSwagFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsZero-shot Language UnderstandingBoolQ, ARC-e, ARC-c, WinoGrande, HellaSwag83.08ARC-e Accuracy8Natural Language ReasoningBoolQ, ARC-e, ARC-c, WinoGrande (WinoG), HellaSwag (HelloS)75.2BoolQ Accuracy4
Natural Language ReasoningBoolQ, ARC-e, ARC-c, WinoGrande (WinoG), HellaSwag (HelloS)75.2BoolQ Accuracy4