Evidence-Grounded Claim-Level AI Hallucination Detector

This paper introduced an evidence-grounded system for assessing the factual reliability of LLM-generated text. It brings together text preprocessing, claim extraction, named entity recognition, query generation, external evidence retrieval, claim-evidence matching, hallucination scoring and explanation generation. In contrast to a plain response-level classifier, it evaluates factual claims one at a time, so a response can contain both supported and problematic statements without losing that distinction. The implementation shows that hallucination verification can be built into a practical application based on Python and FastAPI. Initial testing gave 85% accuracy, 84% precision, 87% recall and an 87% F1-score, along with 85% evidence coverage and about five seconds of average verification time. These numbers are encouraging, but they also underline the need for larger and more rigorous benchmark-based evaluation. The work builds on established research on fine-grained factual evaluation, black-box consistency checking, hallucination benchmarks and retrieval-based verification [1]-[5]. It carries those ideas into a modular, user-facing verification pipeline that puts evidence and interpretability first.

[1]     Z. Ji, N. Lee, R. M. Frieske, T. Yu, D. Su, Y. Xu, E. Ishii, Y. J. Bang, A. Madotto, and P. Fung, “Survey of Hallucination in Natural Language Generation,” ACM Computing Surveys, vol. 55, no. 12, Article 248, pp. 1–38, 2023, doi: 10.1145/3571730. [2]      S. Min, K. Krishna, X. Lyu, M. Lewis, W.-t. Yih, P. Koh, M. Iyyer, L. Zettlemoyer, and H. Hajishirzi, “FActScore: Fine-grained Atomic Evaluation of Factual Precision in Long Form Text Generation,” in Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pp. 12076–12100, 2023, doi: 10.18653/v1/2023.emnlp-main.741. [3]     P. Manakul, A. Liusie, and M. Gales, “SelfCheckGPT: Zero-Resource Black-Box Hallucination Detection for Generative Large Language Models,” in Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pp. 9004–9017, 2023, doi: 10.18653/v1/2023.emnlp-main.557. [4]      C. Niu, Y. Wu, J. Zhu, S. Xu, K. Shum, R. Zhong, J. Song, and T. Zhang, “RAGTruth: A Hallucination Corpus for Developing Trustworthy Retrieval-Augmented Language Models,” in Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp. 10862–10878, 2023, doi: 10.18653/v1/2023.acl-long.585. [5]      J. Li, X. Cheng, X. Zhao, J.-Y. Nie, and J.-R. Wen, “HaluEval: A Large-Scale Hallucination Evaluation Benchmark for Large Language Models,” in Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pp. 6449–6464, 2023, doi: 10.18653/v1/2023.emnlp-main.397. [6]     S. Lin, J. Hilton, and O. Evans, “TruthfulQA: Measuring How Models Mimic Human Falsehoods,” in Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pp. 3214–3252, 2023, doi: 10.18653/v1/2023.acl-long.229. [7]     J. Wei, C. Yang, X. Song, Y. Lu, N. Hu, J. Huang, D. Tran, D. Peng, R. Liu, D. Huang, C. Du, and Q. V. Le, “Long-form factuality in large language models,” arXiv preprint arXiv:2403.18802, 2024. [8]     Y. Song, Y. Kim, and M. Iyyer, “VeriScore: Evaluating the factuality of verifiable claims in long-form text generation,” in Findings of the Association for Computational Linguistics: EMNLP 2024, pp. 9447–9474, 2024, doi: 10.18653/v1/2024.findings-emnlp.552. [9]      D. Muhlgay, O. Ram, I. Magar, Y. Levine, N. Ratner, Y. Belinkov, O. Abend, K. Leyton-Brown, A. Shashua, and Y. Shoham, “Generating Benchmarks for Factuality Evaluation of Language Models,” in Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics (Volume 1: Long Papers), pp. 49–66, 2024, doi: 10.18653/v1/2024.eacl-long.4. [10]      C.-H. Chiang and H.-y. Lee, “Merging Facts, Crafting Fallacies: Evaluating the Contradictory Nature of Aggregated Factual Claims in Long-Form Generations,” in Findings of the Association for Computational Linguistics: ACL 2025, pp. 2734–2751, 2025, doi: 10.18653/v1/2025.findings-acl.160.