ResearchTasksToxic Example RetrievalToxic Example RetrievalFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedUltraChat 8K XSTest-response-Hom variantLESS0.756AUPRC5Jun 12, 2026UltraChat 8K XSTest-response-Het variantLESS70.75AUPRC5Jun 12, 2026