提出新指标与对齐方法，显著提升大模型在RAG中的可信度与拒答能力。
Measuring and Enhancing Trustworthiness of LLMs in RAG through Grounded Attributions and Learning to Refuse
- 设计信任度评分（Trust-Score）量化大模型在RAG中的可信表现
- 提出Trust-Align方法，使26/27模型在多个数据集上性能大幅提升
- 适用于多种开源模型，增强拒答与引用质量，适合实际部署场景
大语言模型是检索增强生成（RAG）系统的核心组件。尽管现有研究多关注端到端RAG系统的整体质量，但对大模型是否适配RAG任务仍缺乏深入评估。为此，本文提出一种综合指标Trust-Score，用于衡量大模型在RAG框架中的可信度。实验表明，多种提示方法（如上下文学习）未能有效提升模型的可信度。因此，我们提出Trust-Align方法以优化模型表现。在ASQA、QAMPARI和ELI5三个数据集上，26/27个经对齐的模型均显著优于基线；以LLaMA-3-8b为例，其在ASQA上提升12.56分，QAMPARI提升36.04分，ELI5提升17.69分。此外，该方法还显著增强了模型正确拒答与提供高质量引用的能力。Trust-Align在不同规模的开源模型（包括LLaMA系列1b至8b、Qwen-2.5系列0.5b至7b及Phi3.5 3.8b）中均表现良好。代码已开源：https://github.com/declare-lab/trust-align。
原文摘要 · Abstract (English)
LLMs are an integral component of retrieval-augmented generation (RAG) systems. While many studies focus on evaluating the overall quality of end-to-end RAG systems, there is a gap in understanding the appropriateness of LLMs for the RAG task. To address this, we introduce Trust-Score, a holistic metric that evaluates the trustworthiness of LLMs within the RAG framework. Our results show that various prompting methods, such as in-context learning, fail to effectively adapt LLMs to the RAG task as measured by Trust-Score. Consequently, we propose Trust-Align, a method to align LLMs for improved Trust-Score performance. 26 out of 27 models aligned using Trust-Align substantially outperform competitive baselines on ASQA, QAMPARI, and ELI5. Specifically, in LLaMA-3-8b, Trust-Align outperforms FRONT on ASQA (up 12.56), QAMPARI (up 36.04), and ELI5 (up 17.69). Trust-Align also significantly enhances models' ability to correctly refuse and provide quality citations. We also demonstrate the effectiveness of Trust-Align across different open-weight models, including the LLaMA series (1b to 8b), Qwen-2.5 series (0.5b to 7b), and Phi3.5 (3.8b). We release our code at https://github.com/declare-lab/trust-align.
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。