ResearchTasksHelpfulness AssessmentFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedQualitative Assessment DatasetEXAONE 3.5 32B100Not Overrefuse Rate (Content-safety)4Mar 23, 2026AdvBench-audio SafeRRS86.44BRR3May 8, 2026