ResearchDatasetsDoNotAnswerDoNotAnswerFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsSafety EvaluationDoNotAnswer Framed0HRR96Toxicity and Harmful Content DetectionDoNotAnswer99.98Score5Safety DetectionDoNotAnswer (held-out)97.4AUROC5