让大模型生成更真实长文本，解决幻觉问题。
FactAlign: Long-form Factuality Alignment of Large Language Models
- 采用细粒度句级对齐算法fKTO，提升事实准确性。
- 在开放域任务中，事实准确率显著提升，帮助性不下降。
- 适合需要高可信长文本输出的研究与应用者。
大型语言模型在信息获取方面展现出巨大潜力，但其可靠性受限于幻觉和生成非事实内容的问题，尤其在长篇回答中，评估与确保事实准确性尤为复杂。本文提出FactAlign，一种新型对齐框架，旨在增强大模型长篇回复的事实性，同时保持其有用性。引入fKTO——一种细粒度、句级的对齐算法，扩展了Kahneman-Tversky优化（KTO）方法。利用近期自动事实性评估技术，FactAlign借助细粒度事实评估引导对齐过程。在开放域提示和信息查询任务上的实验表明，FactAlign显著提升了大模型回复的事实准确性，同时改善了其有用性。进一步分析发现，FactAlign可使模型在不牺牲事实精度的前提下提供更多信息，从而提升事实F1分数。源代码、数据集及训练模型已公开于https://github.com/MiuLab/FactAlign。
原文摘要 · Abstract (English)
Large language models have demonstrated significant potential as the next-generation information access engines. However, their reliability is hindered by issues of hallucination and generating non-factual content. This is particularly problematic in long-form responses, where assessing and ensuring factual accuracy is complex. In this paper, we address this gap by proposing FactAlign, a novel alignment framework designed to enhance the factuality of LLMs' long-form responses while maintaining their helpfulness. We introduce fKTO, a fine-grained, sentence-level alignment algorithm that extends the Kahneman-Tversky Optimization (KTO) alignment method. Leveraging recent advances in automatic factuality evaluation, FactAlign utilizes fine-grained factuality assessments to guide the alignment process. Our experiments on open-domain prompts and information-seeking questions demonstrate that FactAlign significantly improves the factual accuracy of LLM responses while also improving their helpfulness. Further analyses identify that FactAlign is capable of training LLMs to provide more information without losing factual precision, thus improving the factual F1 score. Our source code, datasets, and trained models are publicly available at https://github.com/MiuLab/FactAlign
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。