FactCorrector:一种基于图的长文本事实性校正方法
FactCorrector: A Graph-Inspired Approach to Long-Form Factuality Correction of Large Language Models
AI总结:
FactCorrector通过结构化反馈提升大语言模型的事实性准确性,优于现有基线方法。
摘要:
大型语言模型(LLMs)被广泛应用于知识密集型应用,但常常生成事实性错误的响应。一种有前景的纠正方法是通过反馈来纠正LLMs。因此,在本文中,我们引入了FactCorrector,一种新的事后校正方法,能够在不重新训练的情况下跨领域适应,并利用关于原始响应事实性的结构化反馈来生成校正。为了支持对事实性校正方法的严谨评估,我们还开发了VELI5基准,一个包含系统性注入事实性错误和真实校正的新型数据集。在VELI5以及几个流行的长文本事实性数据集上的实验表明,FactCorrector方法在提高事实精度的同时保持了相关性,优于强大的基线方法。我们通过https://ibm.biz/factcorrector发布了我们的代码。
英文摘要:
Large language models (LLMs) are widely used in knowledge-intensive applications but often generate factually incorrect responses. A promising approach to rectify these flaws is correcting LLMs using feedback. Therefore, in this paper, we introduce FactCorrector, a new post-hoc correction method that adapts across domains without retraining and leverages structured feedback about the factuality of the original response to generate a correction. To support rigorous evaluations of factuality correction methods, we also develop the VELI5 benchmark, a novel dataset containing systematically injected factual errors and ground-truth corrections. Experiments on VELI5 and several popular long-form factuality datasets show that the FactCorrector approach significantly improves factual precision while preserving relevance, outperforming strong baselines. We release our code at https://ibm.biz/factcorrector.
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。