当上下文产生误导时：用于稳健检索增强生成的意图引导解码
When Context Misleads: Intent-Guided Decoding for Robust Retrieval-Augmented Generation
浏览论文内容
中文总结 AI 辅助
针对现有RAG系统固定信任策略的缺陷，提出意图引导解码（IGD）框架，经多基准评估，可显著提升事实恢复能力，同时保持或改善上下文遵循行为。
中文摘要 AI 辅助
检索增强生成（RAG）通过将生成过程锚定在外部证据上提升大语言模型，但也引入了来源信任问题：检索到的上下文可能有用、无关甚至具有误导性。现有RAG系统对检索证据采用固定信任策略，要么过度信任错误上下文，要么在用户明确要求遵循上下文时未能充分利用上下文。因此，我们提出意图引导解码（Intent-Guided Decoding，IGD）框架，该框架根据用户意图在检索上下文和参数记忆之间进行仲裁。IGD采用答案级过滤和令牌级校正，引导最终解码轨迹在检索上下文与参数记忆之间调整。我们在五个大语言模型上的三个忠实问答基准和三个事实冲突基准上评估IGD，结果显示IGD显著提升事实恢复能力，在事实冲突基准上相比直接RAG实现了高达65.4个百分点的提升，同时保持或改善了严格的上下文遵循行为，这些发现凸显了在RAG中平衡事实性与忠实性的重要性。
英文摘要
Retrieval-augmented generation (RAG) improves large language models by grounding generation in external evidence, but it also introduces a source trust problem: retrieved context may be useful, irrelevant, or even misleading. Existing RAG systems often apply a fixed trust policy toward retrieved evidence, which can either over-trust incorrect context or underuse context when the user explicitly asks for context-following behavior. Therefore, we propose Intent-Guided Decoding (IGD), a framework that arbitrates between retrieved context and parametric memory according to user intent. IGD uses answer-level filtering and token-level correction to steer the final decoding trajectory between retrieved context and parametric memory. We evaluate IGD on three faithful QA benchmarks and three factual-conflict benchmarks across five LLMs, IGD substantially improves factual recovery, achieving gains of up to 65.4 percentage points on factual-conflict benchmarks over Direct RAG, while preserving or improving strict context-following behavior, this findings highlight the importance of balancing factuality and faithfulness in RAG.
