ResearchTasksCorrectness detectionCorrectness detectionFollowBenchmarksDatasetSOTA methodSortMost resultsRecently updatedMost papersApplyDatasetSOTA methodMetricTrendResultsLast updatedTruthfulQALlama-3B0.97AUC30Mar 18, 2026Non-AmbigQACredal79.32AUROC20Mar 12, 2026Non-MAQACredal77.53AUROC20Mar 12, 2026MMLU ProCredal87.98AUROC20Mar 12, 2026SciQ, CSQA, FEVER average (cross-dataset)Qwen2-7B0.69AUC6Feb 26, 2026TruthfulQA (in-domain)Llama-1B0.924AUC6Feb 26, 2026