arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2503.09153cs.CLcs.AI
收藏

LLM的幻觉可用吗?基于LLM的负向推理用于虚假新闻检测

Is LLMs Hallucination Usable? LLM-based Negative Reasoning for Fake News Detection

Yangzhou University(扬州大学);Auburn University(奥本大学);The Education University of Hong Kong(香港教育大学)

Chaowei Zhang, Zongling Feng, Zewei Zhang, Jipeng Qiang, Guandong Xu, Yun Li

AI总结:

本文提出利用LLM幻觉生成负向推理,并通过自强化推理校正方法SR$^3$和模型NRFE/NRFE-D学习新闻与推理间的语义一致性,在三个虚假新闻数据集上验证了其优于现有基线方法。

摘要:

由知识幻觉引发的可疑响应可能导致LLM在决策中的能力不稳定。然而,LLM的幻觉是否可能被用于生成负向推理以促进虚假新闻检测,此前从未被研究过。本研究提出了一种新颖的有监督自强化推理校正方法——SR$^3$,该方法通过LLM反思为新闻同时生成常见的合理推理和错误理解(负向推理),用于语义一致性学习。在此基础上,我们构建了一个基于负向推理的新闻学习模型——NRFE,该模型利用正向或负向的新闻-推理对来学习它们之间的语义一致性。为避免标签隐含推理的影响,我们部署了一个学生模型——NRFE-D,该模型仅以新闻内容作为输入,通过从NRFE中蒸馏知识来检验我们方法的性能。在三个流行的虚假新闻数据集上验证的实验结果表明,与三类基线方法(包括在LLM上的提示、在预训练SLM上的微调以及其他代表性的虚假新闻检测方法)相比,我们的方法具有优越性。

英文摘要:

The questionable responses caused by knowledge hallucination may lead to LLMs' unstable ability in decision-making. However, it has never been investigated whether the LLMs' hallucination is possibly usable to generate negative reasoning for facilitating the detection of fake news. This study proposes a novel supervised self-reinforced reasoning rectification approach - SR$^3$ that yields both common reasonable reasoning and wrong understandings (negative reasoning) for news via LLMs reflection for semantic consistency learning. Upon that, we construct a negative reasoning-based news learning model called - \emph{NRFE}, which leverages positive or negative news-reasoning pairs for learning the semantic consistency between them. To avoid the impact of label-implicated reasoning, we deploy a student model - \emph{NRFE-D} that only takes news content as input to inspect the performance of our method by distilling the knowledge from \emph{NRFE}. The experimental results verified on three popular fake news datasets demonstrate the superiority of our method compared with three kinds of baselines including prompting on LLMs, fine-tuning on pre-trained SLMs, and other representative fake news detection methods.

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。

↑