ResearchTasksNear-duplicate identificationFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedFull C4 (365M documents)SEDD27,207,144Number of Duplicates2May 19, 2026C4 Sampled 50M documentsSEDD2,766,436Number of Duplicates2May 19, 2026RealNews 30M documentsSEDD5,426,172Number of Duplicates2May 19, 2026