Authors

Bilal Khan

Assistant Professor, Project Management, Department of Management, India

Abstract

Retrieval-Augmented Generation has become an important approach for improving the factual reliability of large language models by providing external evidence during response generation. However, retrieval augmentation does not guarantee that every generated statement is supported by the retrieved material. Large language models can still produce unsupported, partially supported, or contradictory claims even when relevant evidence is available in the retrieval context. This problem is particularly important for systems expected to provide reliable factual answers because an otherwise accurate response may contain a small number of incorrect claims that are difficult to identify through response-level evaluation. This article examines evidence-grounded hallucination detection and correction using claim-level verification. The study develops a conceptual framework in which generated responses are decomposed into individual factual claims, each claim is matched against retrieved evidence, and unsupported or contradictory claims are subsequently subjected to evidence-based correction. The framework is informed by recent research on retrieval-augmented generation, factuality evaluation, hallucination detection, evidence attribution, and automated verification. Particular attention is given to evidence-bound factual repair approaches that separate semantic localization of factual errors from deterministic realization of evidence-grounded repairs. Existing research including FActScore, SelfCheckGPT, RAGTruth, RAGChecker, RAGAs, ARES, is integrated to establish the theoretical basis of the proposed approach. The analysis indicates that claim-level verification provides a more granular mechanism for identifying factual weaknesses than response-level evaluation, while evidence-constrained correction can reduce the risk of introducing new unsupported information during repair. The proposed framework provides a foundation for developing RAG systems that not only retrieve evidence but also verify, localize, correct, and transparently attribute individual factual claims.

Keywords

Retrieval-augmented generation hallucination detection claim-level verification factuality evidence grounding large language models factual correction RAG evidence attribution trustworthy artificial intelligence

Citation of this Article

Bilal Khan. (2026). Evidence-Grounded Hallucination Detection and Correction in Retrieval-Augmented Large Language Models Using Claim-Level Verification. International Current Journal of Engineering and Science (ICJES), 5(9), 19-24. Article DOI: https://doi.org/10.47001/ICJES/2026.509003 

Licence Copyright (c) 2026 International Current Journal of Engineering and Science. This work is licensed under a Creative Commons Attribution Non Commercial 4.0 International Licence.

References

  1. Dhuliawala, S., Komeili, M., Xu, J., Raileanu, R., Li, X., Celikyilmaz, A., & Weston, J. (2024). Chain-of-verification reduces hallucination in large language models. Findings of the Association for Computational Linguistics: ACL 2024, 3563–3578. https://doi.org/10.18653/v1/2024.findings-acl.212
  2. Es, S., James, J., Espinosa-Anke, L., & Schockaert, S. (2024). RAGAs: Automated evaluation of retrieval augmented generation. In Proceedings of the 18th Conference of the European Chapter of the Association for Computational Linguistics: System Demonstrations (pp. 150–158). Association for Computational Linguistics. https://doi.org/10.18653/v1/2024.eacl-demo.16
  3. Ge, Z., Wu, Y., Chin, D. W. K., Lee, R. K.-W., & Cao, R. (2025). Resolving conflicting evidence in automated fact-checking: A study on retrieval-augmented LLMs. In Proceedings of the Thirty-Fourth International Joint Conference on Artificial Intelligence (pp. 9656–9664). https://doi.org/10.24963/ijcai.2025/1073
  4. Godbole, A., & Jia, R. (2025). Verify with caution: The pitfalls of relying on imperfect factuality metrics. Findings of the Association for Computational Linguistics: ACL 2025, 22889–22912. https://doi.org/10.18653/v1/2025.findings-acl.1175
  5. Lewis, P., Perez, E., Piktus, A., Petroni, F., Karpukhin, V., Goyal, N., Kuttler, H., Lewis, M., Yih, W.-T., Rocktäschel, T., Riedel, S., & Kiela, D. (2020). Retrieval-augmented generation for knowledge-intensive NLP tasks. In Advances in Neural Information Processing Systems (Vol. 33, pp. 9459–9474).
  6. Manakul, P., Liusie, A., & Gales, M. J. F. (2023). 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). Association for Computational Linguistics. https://doi.org/10.18653/v1/2023.emnlp-main.557
  7. Agha, C. W., Mokwenye, I. I., Atikpo, E., & Nwulu, N. (2025). Gumbel Distribution-Based Flood Frequency Analysis for Coastal Flood Risk Assessment in Ugborodo Community, Niger Delta, Nigeria. Environmental Technology and Science Journal, 16(1), 133-143. https://doi.org/10.4314/etsj.v16i1.13
  8. Min, S., Krishna, K., Lyu, X., Lewis, M., Yih, W.-T., Koh, P. W., Iyyer, M., Zettlemoyer, L., & Hajishirzi, H. (2023). 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). Association for Computational Linguistics. https://doi.org/10.18653/v1/2023.emnlp-main.741
  9. Niu, C., Wu, Y., Zhu, J., Xu, S., Shum, K.-S., Zhong, R., Song, J., & Zhang, T. (2024). 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). Association for Computational Linguistics. https://doi.org/10.18653/v1/2024.acl-long.585
  10. Rajendran, S., & Singaravelu, D. (2026). Evidence-bound factual repair in retrieval-augmented LLM answers: Separating semantic localization from deterministic realization (Version 1) [Preprint]. Preprints.org. https://doi.org/10.20944/preprints202609.0490.v1
  11. Ru, D., Qiu, L., Hu, X., Zhang, T., Shi, P., Chang, S., Jiayang, C., Wang, C., Sun, S., Li, H., Zhang, Z., Wang, B., Jiang, J., He, T., Wang, Z., Liu, P., Zhang, Y., & Zhang, Z. (2024). RAGChecker: A fine-grained framework for diagnosing retrieval-augmented generation. arXiv. https://arxiv.org/abs/2408.08067
  12. Russo, D., Menini, S., Staiano, J., & Guerini, M. (2025). Face the facts! Evaluating RAG-based pipelines for professional fact-checking. In Proceedings of the 18th International Natural Language Generation Conference (pp. 846–865). Association for Computational Linguistics. https://aclanthology.org/2025.inlg-main.50/
  13. Saad-Falcon, J., Khattab, O., Potts, C., & Zaharia, M. (2024). ARES: An automated evaluation framework for retrieval-augmented generation systems. In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers). Association for Computational Linguistics. https://aclanthology.org/2024.naacl-long.20/