ResearchTasksMulti-turn Safety Risk AssessmentMulti-turn Safety Risk AssessmentFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedMulti-turn Safety 100 random sampled tasksToolShield0.15ASR8Feb 26, 2026PostgreSQLw/o Defense1ASR8Feb 26, 2026Terminal tasksw/o Defense96ASR8Feb 26, 2026Playwright tasksw/o Defense80ASR8Feb 26, 2026Filesystemw/o Defense96ASR8Feb 26, 2026