ResearchDatasetsCSR benchmarksFollowBenchmarksTask NameDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyTask NameDataset NameSOTA ResultTrendResultsScaling behavior extrapolationCSR benchmarks Downstream & Upstream HellaSwag ARC PIQA WinoGrande OpenBookQA SIQA BoolQ (test)88.89Best Task Percentage6
Scaling behavior extrapolationCSR benchmarks Downstream & Upstream HellaSwag ARC PIQA WinoGrande OpenBookQA SIQA BoolQ (test)88.89Best Task Percentage6