ResearchDatasetsHellaSwag, WinoGrande, IFEval, MMLUHellaSwag, WinoGrande, IFEval, MMLUFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsPreservation of General CapabilitiesHellaSwag, WinoGrande, IFEval, MMLU20.4HellaSwag Delta44