ResearchTasksResponse-side Safety ClassificationFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedResponse-side Safety Benchmarks HarmB, SafeRLHF, BeaverTails, XSTest, Aegis2, WildG, ThinkQwen3Guard-0.8B-Gen98.5HarmB Score12Jul 8, 2026
Response-side Safety Benchmarks HarmB, SafeRLHF, BeaverTails, XSTest, Aegis2, WildG, ThinkQwen3Guard-0.8B-Gen98.5HarmB Score12Jul 8, 2026