ResearchDatasetsLogiHardFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsLogical ReasoningLogiHard-2k Hard-Mode IRT-CAT79.5H-Comb Accuracy13Logical ReasoningLogiHard-2k Full-set Baseline84Original Accuracy13