通过改写输入和动态聚类，提升问答中幻觉检测的准确性与稳定性。
Semantic Reformulation Entropy for Robust Hallucination Detection in QA Tasks
- 用语义改写扩展输入，减少模型表面倾向带来的偏差。
- 采用能量驱动的渐进聚类，使语义分组更稳定。
- 在SQuAD和TriviaQA上优于基线，适合高可靠性问答场景。
大语言模型在问答任务中面临幻觉问题，即因认知不确定性产生流畅但事实错误的回答。现有基于熵的语义级不确定性估计方法受限于采样噪声和可变长度答案的不稳定的聚类。本文提出语义改写熵（SRE），从两方面改进：首先，通过输入侧的语义改写生成忠实的同义表达，扩大估计空间，降低由解码器表面倾向带来的偏差；其次，采用渐进式的能量驱动混合聚类，增强语义分组的稳定性。在SQuAD和TriviaQA上的实验表明，SRE显著优于强基线，提供更鲁棒且泛化性更强的幻觉检测能力。结果表明，结合输入多样化与多信号聚类能显著提升语义级不确定性估计效果。
原文摘要 · Abstract (English)
Reliable question answering with large language models (LLMs) is challenged by hallucinations, fluent but factually incorrect outputs arising from epistemic uncertainty. Existing entropy-based semantic-level uncertainty estimation methods are limited by sampling noise and unstable clustering of variable-length answers. We propose Semantic Reformulation Entropy (SRE), which improves uncertainty estimation in two ways. First, input-side semantic reformulations produce faithful paraphrases, expand the estimation space, and reduce biases from superficial decoder tendencies. Second, progressive, energy-based hybrid clustering stabilizes semantic grouping. Experiments on SQuAD and TriviaQA show that SRE outperforms strong baselines, providing more robust and generalizable hallucination detection. These results demonstrate that combining input diversification with multi-signal clustering substantially enhances semantic-level uncertainty estimation.
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。