ResearchTasksPolicy FollowingPolicy FollowingFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedPOLICYSHIFTBENCH (Adaptive split)Human Performance89Accuracy44Jul 8, 2026POLICYSHIFTBENCH ShiftHuman Performance90Accuracy22Jul 8, 2026