ResearchTasksLatent ReasoningLatent ReasoningFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedGSM8K (test)GTS1.84SG Score8Feb 26, 2026partitioned-k=2 query-blind (100 problems)BagMerge90Accuracy6Jul 3, 2026partitioned-k=2 query-known (100 problems)CanonicalMerge94Accuracy6Jul 3, 2026