大语言模型中的几何感知幻觉检测
Geometry-Aware Hallucination Detection in Large Language Models
AI总结：
提出GA-ICL框架，利用冻结LLM的潜在表示建模局部流形和类别原型几何，选择上下文示例以检测幻觉，在FEVER和HaluEval基准上优于基线方法。
摘要：
大型语言模型（LLM）经常生成事实不正确或未经支持的内容，通常称为幻觉。先前的工作探索了解码策略、检索增强和监督微调用于幻觉检测，而最近的研究表明，上下文学习（ICL）可以显著影响事实可靠性。然而，现有的ICL示例选择方法通常依赖于表面相似性启发式方法，并且在任务和模型上表现出有限的鲁棒性。我们提出GA-ICL，一种几何感知的示例采样框架，用于选择上下文示例，该框架利用从冻结LLM中提取的潜在表示。通过联合建模局部流形结构和类别感知的原型几何，GA-ICL根据示例与学习原型的接近程度进行选择，而不仅仅是基于词汇或嵌入相似性。在事实验证（FEVER）和幻觉检测（HaluEval）基准上，GA-ICL在大多数评估设置中优于标准ICL选择基线，在对话和摘要任务上尤其有显著提升。该方法在温度扰动和模型变化下保持鲁棒性，表明与启发式检索策略相比具有更高的稳定性。虽然在较小模型规模下的某些问答场景中，词汇检索仍可能具有竞争力，但我们的结果表明，几何感知的原型选择为幻觉检测提供了一种可靠且训练轻量的方法，无需修改LLM参数。在Phi-14B和Qwen3-32B上的扩展评估证实，GA-ICL能有效扩展到更大模型，在包括较小模型显示边界条件限制的问答任务在内的所有比较基线上均表现优异，为改进ICL示例选择提供了原则性方向。
英文摘要：
Large language models (LLMs) frequently generate factually incorrect or unsupported content, commonly referred to as hallucinations. Prior work has explored decoding strategies, retrieval augmentation, and supervised fine-tuning for hallucination detection, while recent studies show that in-context learning (ICL) can substantially influence factual reliability. However, existing ICL demonstration selection methods often rely on surface-level similarity heuristics and exhibit limited robustness across tasks and models. We propose GA-ICL, a geometry-aware demonstration sampling framework for selecting in-context demonstrations that leverages latent representations extracted from frozen LLMs. By jointly modeling local manifold structure and class-aware prototype geometry, GA-ICL selects demonstrations based on their proximity to learned prototypes rather than lexical or embedding similarity alone. Across factual verification (FEVER) and hallucination detection (HaluEval) benchmarks, GA-ICL outperforms standard ICL selection baselines in the majority of evaluated settings, with particularly strong gains on dialogue and summarization tasks. The method remains robust under temperature perturbations and model variation, indicating improved stability compared to heuristic retrieval strategies. While lexical retrieval can remain competitive in certain question-answering regimes at smaller model scales, our results demonstrate that geometry-aware prototype selection provides a reliable and training-light approach for hallucination detection without modifying LLM parameters. Extended evaluations on Phi-14B and Qwen3-32B confirm that GA-ICL scales effectively to larger models, outperforming all compared baselines including on QA tasks where smaller models show boundary-condition limitations, offering a principled direction for improved ICL demonstration selection.
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。
