ResearchDatasetsGemini-annotatedFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsMachine Translation RankingGemini-annotated held-out Zh↔En (test)82.62Accuracy10Multi-label Value ClassificationGemini-annotated N=4,000 (test)0.752Self-direction Score1