ResearchDatasetsMM-AlignBenchFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsHuman Preference AlignmentMM-AlignBench 1.0 (test)84.9Win Rate18Multi-modal preference alignmentMM-AlignBench62.3Winning Rate6