ResearchTasksPairwise Preference ComparisonFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedPairwise Preference Comparisons 1.5B Scale (test)STACKELBERGGDA FOLLOWER0.834Avg Preference Score30Feb 26, 2026Qwen2.5-3B responses (test)NASH-MD-PG82.7Avg Preference Score30Feb 26, 2026150 prompt-response pairsRL-trained policy63.333Win Rate9May 29, 2026HH-RLHF held-out (test)DP-RLHF53.02Win Rate6Mar 25, 2026
Pairwise Preference Comparisons 1.5B Scale (test)STACKELBERGGDA FOLLOWER0.834Avg Preference Score30Feb 26, 2026