arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2504.10168cs.CLcs.AI

HalluSearch 在 SemEval-2025 任务 3 中的应用:用于幻觉检测的搜索增强型 RAG 流水线

HalluSearch at SemEval-2025 Task 3: A Search-Enhanced RAG Pipeline for Hallucination Detection

School of Information Technology, Newgiza University(新吉萨大学信息技术学院)

Mohamed A. Abdallah, Samhaa R. El-Beltagy

AI总结:

本文提出多语言流水线HalluSearch,结合检索增强验证与细粒度事实拆分检测14种语言LLM输出中的伪造文本片段,在英语和捷克语任务中排名第四,为多语言幻觉检测提供有效方案并指出需优化低覆盖语言表现。

摘要:

本文提出HalluSearch,这是一个多语言流水线,旨在检测大型语言模型(LLM)输出中的伪造文本片段。该流水线是Mu-SHROOM(关于幻觉及相关可观察过度生成错误的多语言共享任务)的一部分,将检索增强验证与细粒度事实拆分相结合,以识别和定位14种不同语言中的幻觉。实证评估表明,HalluSearch表现具有竞争力,在英语(前10%以内)和捷克语中均排名第四。虽然该系统的基于检索的策略总体上被证明是稳健的,但它在在线覆盖有限的语言中面临挑战,这凸显了进一步研究的必要性,以确保在不同语言环境中一致的幻觉检测。

英文摘要:

In this paper, we present HalluSearch, a multilingual pipeline designed to detect fabricated text spans in Large Language Model (LLM) outputs. Developed as part of Mu-SHROOM, the Multilingual Shared-task on Hallucinations and Related Observable Overgeneration Mistakes, HalluSearch couples retrieval-augmented verification with fine-grained factual splitting to identify and localize hallucinations in fourteen different languages. Empirical evaluations show that HalluSearch performs competitively, placing fourth in both English (within the top ten percent) and Czech. While the system's retrieval-based strategy generally proves robust, it faces challenges in languages with limited online coverage, underscoring the need for further research to ensure consistent hallucination detection across diverse linguistic contexts.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑