ResearchTasksMalicious Goal Attack (Longer Token Generation)For AI assistantsMarkdownContent indexMalicious Goal Attack (Longer Token Generation)FollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedPKU-SafeRLHF (test)RankPoison50.17RM Length Accuracy3Feb 26, 2026