LLMQuoter：通过从大上下文中高效提取引文增强 RAG 能力
LLMQuoter: Enhancing RAG Capabilities Through Efficient Quote Extraction From Large Contexts
AI总结：
本文提出轻量级蒸馏模型 LLMQuoter，通过 LoRA 微调与“先引文后回答”策略从大上下文中提取关键证据，显著提升大小模型的 RAG 推理准确率。
摘要：
我们提出 LLMQuoter，这是一个轻量级、基于蒸馏的模型，旨在通过为下游推理任务提取最相关的文本证据来增强检索增强生成（RAG）。LLMQuoter 基于 LLaMA-3B 架构构建，并在 HotpotQA 的 15,000 个样本子集上使用低秩适应（LoRA）进行微调，采用“先引文后回答”策略，在将精选片段传递给推理模型之前高效识别关键引文。该工作流程降低了认知开销，并优于检索增强微调（RAFT）等全上下文方法，在小型和大型语言模型上均实现超过 20 个百分点的准确率提升。通过利用来自高性能教师模型的知识蒸馏，LLMQuoter 在资源高效的微调设置中取得了具有竞争力的结果。它使先进 RAG 能力更加普及，无需大量模型重训即可带来显著性能提升。我们的结果凸显了蒸馏式引文推理在简化复杂工作流程方面的潜力，为研究人员和从业者提供了可扩展且实用的解决方案。
英文摘要：
We introduce LLMQuoter, a lightweight, distillation-based model designed to enhance Retrieval Augmented Generation (RAG) by extracting the most relevant textual evidence for downstream reasoning tasks. Built on the LLaMA-3B architecture and fine-tuned with Low-Rank Adaptation (LoRA) on a 15,000-sample subset of HotpotQA, LLMQuoter adopts a "quote-first-then-answer" strategy, efficiently identifying key quotes before passing curated snippets to reasoning models. This workflow reduces cognitive overhead and outperforms full-context approaches like Retrieval-Augmented Fine-Tuning (RAFT), achieving over 20-point accuracy gains across both small and large language models. By leveraging knowledge distillation from a high-performing teacher model, LLMQuoter achieves competitive results in a resource-efficient fine-tuning setup. It democratizes advanced RAG capabilities, delivering significant performance improvements without requiring extensive model retraining. Our results highlight the potential of distilled quote-based reasoning to streamline complex workflows, offering a scalable and practical solution for researchers and practitioners alike.
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。
