ResearchDatasetsCocoNotFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsSafety EvaluationCocoNot0.613Safety Score36Refusal EvaluationCoCoNot Orig96.1Refusal Rate7Over-refusal EvaluationCoCoNot Contrast1.58Over-refusal Rate7