ResearchDatasetsUnsafeDiffUnsafeDiffFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsJailbreak AttackUnsafeDiff1.7Attack Success Rate (ASR)12Safety EvaluationUnsafeDiff (test)94.2F1 Score11Black-box NSFW Filter AttackUnsafeDiff (test)22Adult Bypass Rate2