ResearchDatasetsInsecure codeInsecure codeFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsDishonesty EvaluationInsecure code (test)48.91Benchmark Dishonesty32Data RankingInsecure code0.71AUROC28Coding PerformanceInsecure-code 1000-prompt held-out93.9Task Success Rate7