ResearchTasksJailbreak Attack ResilienceFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedHarmBench (overall)Llama 3.2 (3B)5ASR6May 14, 2026HarmBench Misinfo. & Disinfo.Phi-3 Mini80ASR3May 14, 2026HarmBench General HarmQwen 3 (4B)0Attack Success Rate (ASR)3May 14, 2026