ResearchTasksClosure TestFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast Updatednatural-text batch (nat)Pythia 1B1.95zDelta Loss3Jun 9, 2026Synthetic Induction Batch (synth)OLMo 1B1.83Z-Delta Loss2Jun 9, 2026