ResearchTasksAccess ControlFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedMuses-Bench Deepseek max-turns = 10Harness-MU100Privacy26Jun 23, 202612 template–attack datasets (test)Harness-MU100Privacy8Jun 23, 2026Synthetic D&D dataset (test)Static Filter92Accuracy8Feb 26, 2026TOFUSUDOLM98.13Accuracy5Feb 26, 2026MedicalSUDOLM100Accuracy5Feb 26, 2026Generated access control dataset (test)RBAC100Upload Homework4Feb 26, 2026