@inproceedings{huang-etal-2026-rlseek,
    title = "{RLS}eek: Evidence-Grounded Reasoning for {RAG} Hallucination Detection",
    author = "Huang, Zhaoheng  and
      Wen, Dacheng  and
      Zhu, Yutao  and
      Lian, Xiaoying  and
      Liang, Yushi  and
      Hao, Kai  and
      Li, Nan  and
      Zhang, Liangjie  and
      Zhang, Qi  and
      Wen, Ji-Rong  and
      Dou, Zhicheng  and
      Wu, Fangzhao",
    editor = "Liakata, Maria  and
      Moreira, Viviane P.  and
      Zhang, Jiajun  and
      Jurgens, David",
    booktitle = "Proceedings of the 64th Annual Meeting of the {A}ssociation for {C}omputational {L}inguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2026",
    address = "San Diego, California, United States",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingest-acl/2026.acl-long.1492/",
    pages = "32329--32347",
    ISBN = "979-8-89176-390-6",
    abstract = "Large language models (LLMs) in retrieval-augmented generation systems can still produce hallucinations, generating content that is unsupported or contradicted by the source texts and undermines reliability. Recent work addressed this problem by training span-level hallucination detectors using reinforcement learning (RL) and chain-of-thought (CoT) reasoning. In this work, we show through error analysis that incorrect predictions by existing reasoning-based detectors are strongly associated with CoT processes that lack explicit grounding in source evidence, particularly when verification steps do not quote or verify claims against the retrieved documents. This behaviour contrasts with human verification practices in benchmarks such as RAGTruth, where evidence quotation is a prerequisite for determining hallucinated spans. Motivated by this observation, we propose an evidence-grounded RL framework, namely RLSeek, to explicitly enforce active evidence seeking during CoT reasoning by requiring quotation of relevant source segments at each verification step. Experiments on the RAGTruth and NewsSum dataset demonstrate consistent improvements in hallucination span detection performance, with limited additional reasoning overhead and improved robustness in out-of-domain settings."
}