ResearchTasksDishonesty EvaluationFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedMistake math (test)PRISM44.16Benchmark Dishonesty96May 21, 2026Insecure code (test)PRISM48.91Benchmark Dishonesty32May 21, 2026Mistake medical (test)CONCEPTINF67.52Dishonesty Accuracy32May 21, 2026