ResearchTasksSafety-Utility Trade-off EvaluationFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedS-Eval, ORFuzzSet, and NQ AggregatedLLM-VA86.81F1 Score72Feb 26, 2026AgentHarmTRIAD + Tri-Guard82.79HS Score15Jun 5, 2026Aggregate (Koala, JBench-B, GSM-8k, SQL-1k, OrBench-H, SorryBench, JBench-H, HEX-PHI)Db as Alpaca77.03Average Score6Mar 13, 2026
Aggregate (Koala, JBench-B, GSM-8k, SQL-1k, OrBench-H, SorryBench, JBench-H, HEX-PHI)Db as Alpaca77.03Average Score6Mar 13, 2026