通过语义级推理图检测RAG中的幻觉，提升准确性。
Detecting Hallucinations in Retrieval-Augmented Generation via Semantic-level Internal Reasoning Graph
- 将归因传播扩展到语义层面，构建推理依赖图
- 在RAGTruth和Dolly-15k上优于现有方法
- 适合关注LLM可信性与推理可解释性的研究者
基于大语言模型的检索增强生成（RAG）系统已取得显著进展，能有效减少事实性幻觉，但忠实性幻觉仍存在。以往检测方法或忽略模型内部推理过程，或粗略处理相关特征，导致判别器难以学习。本文提出一种基于语义级内部推理图的忠实性幻觉检测方法。首先将逐层重要性传播算法从词元层面扩展至语义层面，基于归因向量构建内部推理图，提供更真实的语义依赖表示。进一步设计一个基于小型预训练语言模型的通用框架，利用大模型推理中的依赖关系进行训练与幻觉检测，并可通过阈值动态调整正确样本通过率。实验结果表明，该方法在RAGTruth和Dolly-15k数据集上整体性能优于当前最优基线。
原文摘要 · Abstract (English)
The Retrieval-augmented generation (RAG) system based on Large language model (LLM) has made significant progress. It can effectively reduce factuality hallucinations, but faithfulness hallucinations still exist. Previous methods for detecting faithfulness hallucinations either neglect to capture the models' internal reasoning processes or handle those features coarsely, making it difficult for discriminators to learn. This paper proposes a semantic-level internal reasoning graph-based method for detecting faithfulness hallucination. Specifically, we first extend the layer-wise relevance propagation algorithm from the token level to the semantic level, constructing an internal reasoning graph based on attribution vectors. This provides a more faithful semantic-level representation of dependency. Furthermore, we design a general framework based on a small pre-trained language model to utilize the dependencies in LLM's reasoning for training and hallucination detection, which can dynamically adjust the pass rate of correct samples through a threshold. Experimental results demonstrate that our method achieves better overall performance compared to state-of-the-art baselines on RAGTruth and Dolly-15k.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。