ResearchTasksExplainability classificationExplainability classificationFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedToxic-Chat 0124 (test)LEG large65.99Unsafe F130Feb 26, 2026AEGIS 2.0 (test)LEG large79.6Unsafe F127Feb 26, 2026WildGuardMix human-annotated (test)LEG base60.69F1 Score3Feb 26, 2026