ResearchDatasetsSkillHarmSkillHarmFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsSkill-based attack evaluationSkillHarm Self-Mutating Poisoning69.3ASR6Skill-based attack evaluationSkillHarm Fixed-Payload Poisoning86.3ASR6Skill Injection DetectionSkillHarm SMP setting68.8Detection Rate3Skill Injection DetectionSkillHarm FPP setting55.6Detection Rate3