ResearchTasksLanguage-to-Execution GroundingFollowBenchmarksDataset NameSOTA MethodSortMost resultsRecently updatedMost papersApplyDataset NameSOTA methodMetricTrendResultsLast UpdatedLabGuard-Bench stress benchmark (Random 80/20 split)Gold100Tag F14Jul 1, 2026LabGuard-Bench stress benchmark (Source-held-out)LabGuard-Hybrid76.5Tag F13Jul 1, 2026