ResearchTasksArgument RevisionArgument RevisionFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updated100 arguments Human EvaluationGRPOfull82Preference Rank 1 (%)4Apr 15, 2026Human Evaluation Set 200 edit suggestionsGRPOfull4.41Similarity3Apr 15, 2026