ResearchDatasetsHuman-AuthoredFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsMedical AdviceHuman-Authored17.51Bypass Rate7Self-HarmHuman-Authored1.93Non-safe Rate4Self-HarmHuman-Authored13.69Bypass Rate3Safety ClassificationHuman-Authored Self-harm (test)0.07Violation Rate3