ResearchDatasetsPOLICYSHIFTBENCHPOLICYSHIFTBENCHFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsPolicy FollowingPOLICYSHIFTBENCH (Adaptive split)89Accuracy44Policy FollowingPOLICYSHIFTBENCH Shift90Accuracy22Policy-adaptive safety guardrailingPOLICYSHIFTBENCH (Shift)67F1 Score9Policy-adaptive safety guardrailingPOLICYSHIFTBENCH (Adaptive split)86.8F1 Score9