ResearchTasksConversational Policy EnforcementFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedDynaGuard Privacy Synthetic (test)GPT-5-mini98Accuracy10Apr 29, 2026DynaGuard Privacy Human-annotated (test)GPT-5-mini87Accuracy10Apr 29, 2026DynaGuard Repetition Synthetic (test)ft-4.1-nano93Accuracy10Apr 29, 2026DynaGuard Repetition Human-annotated (test)ft-4.1-nano96Accuracy10Apr 29, 2026