ResearchTasksPolicy EnforcementFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedSide ChannelPraline100Utility12Jun 19, 2026ATBenchPraline100Utility12Jun 19, 2026IntercodeMonte Carlo100Utility12Jun 19, 2026Dynamic Policy E-commerce BenchmarkQwen3-32B-Thinking91Recall6Feb 26, 2026PhantomPolicy (human-reviewed trace labels)Policy-in-prompt40.7Risky-case Violation Rate2Apr 15, 2026
PhantomPolicy (human-reviewed trace labels)Policy-in-prompt40.7Risky-case Violation Rate2Apr 15, 2026