ResearchTasksDefending against gradient-based attacksDefending against gradient-based attacksFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedLlama3 AutoDAN Attack (test)Ours-Fakecom-t10.57ASR10Feb 26, 2026Llama3 GCG Attack (test)Ours-Fakecom-t9.61ASR10Feb 26, 2026