ResearchDatasetsmC4FollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsTokenizer EvaluationmC4 unbalanced 30-lang0.03Compression Rate13Tokenization Intrinsic EvaluationmC4 unbalanced 60-lang (test)0.021Compression Rate5Embedding InversionmC4 (val)81.3Token Accuracy3Reinforcement LearningbusyRingMC46.08Time1