ResearchTasksSafety UnlearningFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedBeaverTailsRepSelect0Harmful Response Probability42Jun 17, 2026SAFEERASER (test)KL4.9RA (ASR)4Feb 26, 2026