CUE-R：超越检索增强生成的最终答案
CUE-R: Beyond the Final Answer in Retrieval-Augmented Generation
AI总结：
CUE-R通过干预基于证据项的检索使用痕迹，评估单次检索增强生成中每个证据项的操作效用，揭示证据项对正确性、基础忠实度和置信度误差的影响。
摘要：
随着语言模型从单次回答生成转向多步骤推理，检索并消耗证据中间推理，评估单个检索项的作用变得尤为重要。现有RAG评估通常针对最终答案质量、引用忠实度或答案级归因，但均未直接针对本文研究的干预式、按证据项效用视角。我们引入CUE-R，一种轻量级的干预式框架，用于测量单次RAG中每个证据项的操作效用，使用浅层可观察的检索使用痕迹。CUE-R通过REMOVE、REPLACE和DUPLICATE操作干预单个证据项，然后测量沿三个效用轴（正确性、基于代理的基础忠实度和置信度误差）以及一个痕迹分歧信号的变化。我们还概述了操作证据角色分类法，以解释干预结果。在HotpotQA和2WikiMultihopQA上使用Qwen-3 8B和GPT-5.2进行实验，揭示了一致模式：REMOVE和REPLACE显著损害正确性和基础性，同时产生大的痕迹偏移，而DUPLICATE通常答案冗余但不完全行为中性。零检索控制确认这些影响源于有意义检索的退化。双支持消融进一步表明，多跳证据项可以非加性交互：移除两个支持对性能的影响远大于单个移除。我们的结果表明，仅答案评估忽略了重要的证据效应，干预式效用分析是RAG评估的实用补充。
英文摘要：
As language models shift from single-shot answer generation toward multi-step reasoning that retrieves and consumes evidence mid-inference, evaluating the role of individual retrieved items becomes more important. Existing RAG evaluation typically targets final-answer quality, citation faithfulness, or answer-level attribution, but none of these directly targets the intervention-based, per-evidence-item utility view we study here. We introduce CUE-R, a lightweight intervention-based framework for measuring per-evidence-item operational utility in single-shot RAG using shallow observable retrieval-use traces. CUE-R perturbs individual evidence items via REMOVE, REPLACE, and DUPLICATE operators, then measures changes along three utility axes (correctness, proxy-based grounding faithfulness, and confidence error) plus a trace-divergence signal. We also outline an operational evidence-role taxonomy for interpreting intervention outcomes. Experiments on HotpotQA and 2WikiMultihopQA with Qwen-3 8B and GPT-5.2 reveal a consistent pattern: REMOVE and REPLACE substantially harm correctness and grounding while producing large trace shifts, whereas DUPLICATE is often answer-redundant yet not fully behaviorally neutral. A zero-retrieval control confirms that these effects arise from degradation of meaningful retrieval. A two-support ablation further shows that multi-hop evidence items can interact non-additively: removing both supports harms performance far more than either single removal. Our results suggest that answer-only evaluation misses important evidence effects and that intervention-based utility analysis is a practical complement for RAG evaluation.
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。
