ResearchDatasetsAgentHazardFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsSafety DetectionAgentHazard Strongest90.87Accuracy56Agentic Safety ModerationAgentHazard109ASR5Dataset Diversity and Coverage EvaluationAgentHazard full0.93Goal-text Entropy1Dataset Diversity and Coverage EvaluationAgentHazard 3-app0.976Goal-text Entropy1