ResearchDatasetsHendrycks MATHHendrycks MATHFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsMathematical ReasoningHendrycks-MATH held-out 1,000-question71.7Accuracy (Base)9Mathematical ReasoningHendrycks MATH36.7Peak Accuracy5Math ReasoningHendrycks MATH61.4Pass@13