arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2502.15845cs.CLcs.AI
收藏

不确定时验证:超越黑盒幻觉检测中的自一致性

Verify when Uncertain: Beyond Self-Consistency in Black Box Hallucination Detection

MIT-IBM Watson AI Lab(MIT-IBM沃森人工智能实验室);IBM Research(IBM研究院)

Yihao Xue, Kristjan Greenewald, Youssef Mroueh, Baharan Mirzasoleiman

AI总结:

提出两阶段检测算法,结合自一致性与交叉一致性,在保持高检测性能的同时降低计算成本,并通过核均值嵌入提供理论解释。

摘要:

大型语言模型(LLM)经常产生幻觉,限制了它们在敏感应用中的可靠性。在黑盒设置中,已经提出了几种基于自一致性的幻觉检测技术。我们经验性地表明,这些方法的性能几乎与监督(黑盒)预言机一样好,在该范式内进一步改进的空间有限。为了解决这一局限性,我们探索了目标模型与额外验证器LLM之间的跨模型一致性检查。利用这一额外信息,我们观察到与纯自一致性方法相比,预言机性能有所提升。然后,我们提出了一种预算友好的两阶段检测算法,该算法仅对一部分情况调用验证器模型。它根据自一致性分类器的不确定性区间,在自一致性和交叉一致性之间动态切换。我们通过核均值嵌入的视角提供了一致性幻觉检测方法的几何解释,提供了更深入的理论见解。在问答式幻觉检测基准上的大量实验表明,该方法在保持高检测性能的同时显著降低了计算成本。

英文摘要:

Large Language Models (LLMs) often hallucinate, limiting their reliability in sensitive applications. In black-box settings, several self-consistency-based techniques have been proposed for hallucination detection. We empirically show that these methods perform nearly as well as a supervised (black-box) oracle, leaving limited room for further gains within this paradigm. To address this limitation, we explore cross-model consistency checking between the target model and an additional verifier LLM. With this extra information, we observe improved oracle performance compared to purely self-consistency-based methods. We then propose a budget-friendly, two-stage detection algorithm that calls the verifier model only for a subset of cases. It dynamically switches between self-consistency and cross-consistency based on an uncertainty interval of the self-consistency classifier. We provide a geometric interpretation of consistency-based hallucination detection methods through the lens of kernel mean embeddings, offering deeper theoretical insights. Extensive experiments on QA-style hallucination detection benchmarks show that this approach maintains high detection performance while significantly reducing computational cost.

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。

↑