ResearchDatasetsGRANDGUARDGRANDGUARDFollowBenchmarksTaskDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyTaskDatasetSOTA resultTrendResultsResponse Safety EvaluationGRANDGUARD unsafe prompts89.8Safe Rate10Prompt-level Risk DetectionGRANDGUARD53.9F1 Score5