ResearchTasksMulti-turn Jailbreak EvaluationMulti-turn Jailbreak EvaluationFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedMHJ LLaMA-3.1-70B (test)HMNS95.2ASR (%)5Apr 14, 2026MHJ Phi-3-Medium-14B (test)HMNS91.6ASR5Apr 14, 2026