ResearchTasksRobustness to Jailbreak AttacksFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedAdaptive attackIPO20Harmful Reasoning Ratio3Mar 4, 2026JBB ParaphrasedIPO16.1Harmful Reasoning Ratio3Mar 4, 2026JailbreakV-28KIPO24.5Harmful Reasoning Ratio3Mar 4, 2026