ResearchTasksJudge AccuracyFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedRMB-SafeRLHF (Target)Seed Rubric69Accuracy4Feb 26, 2026RMB-SafeRLHF (Bench)Biased Rubric Search85.6Accuracy4Feb 26, 2026Ultra-Problem (Target)Seed Rubric61.5Accuracy2Feb 26, 2026Ultra-Problem (Bench)Biased Rubric Search73Accuracy2Feb 26, 2026