ResearchTasksHelpfulness-Safety AlignmentFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedOverallGD2PO-Hard5.808Average Score8Jun 16, 2026AlpacaGD2PO-Hard5.633Useful Score8Jun 16, 2026PKU-AlignmentGD2PO-Hard5.497Useful Score8Jun 16, 2026HH-RLHFGD2PO-Hard4.674Useful Score8Jun 16, 2026