ResearchTasksAddition reasoningAddition reasoningFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedSynthetic Reasoning Add L=100 (test)DNC+CUW60.1Accuracy4Feb 26, 2026Synthetic Reasoning Add (L=50) (test)DNC+CUW94.4Accuracy4Feb 26, 2026