ResearchTasksDecision classificationDecision classificationFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedDamage-audit 139-case extended (held-out)B1b84.2Action Accuracy8Jun 11, 2026Damage-audit (40-case held-out)v176.4Action Accuracy3Jun 11, 2026