用图结构建模文本对齐关系，提升大模型回答的可信度
Graph Alignment Topology as an Inductive Bias for Grounding Detection
- 构建参考文本与生成答案的双分图，用GNN学习对齐结构
- 在四个数据集上超越GPT-4o等模型，准确率显著提升
- 适合需要高事实正确性的医疗等关键领域应用
大型语言模型（LLMs）倾向于生成分布上合理的内容，而非显式验证生成命题是否由源文档蕴含。这种归纳偏置虽促进泛化，但未编码回答与参考文本的对齐性。这限制了其在临床决策支持等需严格事实正确性的场景中的应用。现有幻觉检测方法通过检索增强、自一致性或断言验证提升事实性，但通常不直接学习对齐拓扑结构。为此，本文构建参考信息与LLM输出间的对齐二分图，并训练图神经网络（GNN）通过消息传递建模对齐结构。该方法在四个多样化的幻觉检测与问答数据集上取得当前最优表现，全面优于包括GPT-4o在内的所有对比方法。
原文摘要 · Abstract (English)
Large Language Models (LLMs) are optimized to produce distributionally plausible continuations rather than to explicitly verify whether generated propositions are entailed by source documents. This inductive bias enables generalization, but it does not encode whether responses are grounded with respect to a reference. These issues limit the use of LLMs in domains where strict factual correctness is crucial, such as clinical decision support. Existing hallucination detection approaches improve factuality through retrieval augmentation, self-consistency, or claim verification, but generally do not learn directly over alignment topology. To leverage alignment topology as an inductive bias, we construct aligned bipartite graphs between reference information and LLM outputs and train a graph neural network (GNN) to model alignment structure using message passing. The method achieves state-of-the-art results on four diverse hallucination and question-answering datasets, outperforming all compared methods, including foundational LLMs such as GPT-4o.
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。