ResearchDatasetsBeaverTails, ToxicChat, PKU-SafeRLHFFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsHarmfulness detectionBeaverTails, ToxicChat, PKU-SafeRLHF Hard (test)90.2AUROC63