ResearchDatasetsCSTM-BenchCSTM-BenchFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsMulti-stage Attack DetectionCSTM-Bench (cross-session shard)73.1Detection Rate3Multi-stage Attack DetectionCSTM-Bench (dilution shard)100Detection Rate3