ResearchDatasetsToolEmu and SafeToolBenchToolEmu and SafeToolBenchFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsSecure tool generalizationToolEmu and SafeToolBench Balanced OOD transfer56.9GUARDEDJOINT Score7External Attack Transfer DetectionToolEmu and SafeToolBench (external harmful episodes)99.52Recall @ 5% ID-FPR4
External Attack Transfer DetectionToolEmu and SafeToolBench (external harmful episodes)99.52Recall @ 5% ID-FPR4