通过输入证据对齐，定位大模型生成中的幻觉片段。
Hallucination Span Detection with Input-Side Evidence Alignment
- 基于输入可预测性区分真实与幻觉输出片段。
- 在多个数据集上准确识别幻觉段落，对齐率超80%。
- 适合需要可解释生成结果的研究者与开发者。
大语言模型在条件文本生成中仍受幻觉问题困扰。现有方法多评估整个生成文本的事实性，难以定位具体幻觉片段及其与输入的关系。本文提出“输入侧证据对齐的幻觉片段检测”任务，联合识别幻觉段落并建立输出与输入证据的对齐关系。核心思路是：忠实输出片段可从输入中预测，而幻觉片段则不可预测。我们训练一个编码器模型，从输入表示中预测被掩码的输出片段，并利用预测置信度进行幻觉检测，同时自然生成输入对齐。实验表明，该方法能有效检测幻觉段落并识别有意义的输入证据。人工评估验证了预测对齐的质量。
原文摘要 · Abstract (English)
Hallucinations remain a major obstacle to the reliable use of large language models (LLMs) in conditional text generation. Existing methods primarily assess the factuality of an entire generated text, providing limited insight into which output spans are hallucinated or how they relate to the input. We introduce the task of hallucination span detection with input-side evidence alignment, which jointly identifies hallucinated spans and aligns output tokens with the corresponding input evidence. Our approach is based on the observation that faithful output tokens are predictable from the input, whereas hallucinated tokens are not. We therefore train an encoder-based model to predict masked output tokens from the input representation, using prediction confidence for hallucination detection while naturally producing alignments to the input. Experiments show that the proposed method effectively detects hallucinated spans and identifies meaningful input-side evidence. Human evaluation confirms the quality of the predicted alignments.
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。