ResearchDatasetsHarmful BenchmarksHarmful BenchmarksFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsSafety EvaluationHarmful Benchmarks (CATQA, HEX-PHI, Salad-Base)99.94CATQA Score24Safety AlignmentHarmful Benchmarks Aggregated (test)3.63Harmful Response Rate5