ResearchTasksLong-form reasoningLong-form reasoningFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedAlpacaProbability-1.577Avg LogProb per Answer14Feb 26, 2026NuminaProof (test)SFT (no CoT)-1.017Avg LogProb (Per Answer)14Feb 26, 2026