ResearchTasksDiscovery ManipulationDiscovery ManipulationFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedClawHubBeam-Search Based Attack56Top-3 Accuracy5May 13, 2026ClawHub 0-dayBeam-Search Based Attack94Win-Rate2May 13, 2026ClawHub average-dayBeam-Search Based Attack74.14Win-Rate1May 13, 2026