ResearchDatasetsBeaverTails, ToxicChat, PKU-SafeRLHFBeaverTails, ToxicChat, PKU-SafeRLHFFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsHarmfulness detectionBeaverTails, ToxicChat, PKU-SafeRLHF Hard (test)90.2AUROC63