ResearchDatasetsGreater-Than taskFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsHead RecoveryGreater-Than task Pythia-410M (N=200)80.4Hit Rate @55Head RecoveryGreater-Than task GPT-2 (N=200)83.2Accuracy @55