ResearchTasksEmbodied Agent Planning (Adversarial Safety Evaluation)FollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedSafeAgentBench Unsafe Tasks - JailbreakLota-Bench54.85SR12May 12, 2026