ResearchTasksResponse-level Hallucination DetectionResponse-level Hallucination DetectionFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedRAGTruth (test)GASP-threshold (default)74.5AUC15Jul 7, 2026RAGognizeRAGognizer94.91AUROC13Apr 20, 2026RAGTruth QALettuceDetect-L91.89AUROC13Apr 20, 2026HDM-BenchHallucinationProbes76.3AUROC11Apr 20, 2026