CUE-R: Beyond the Final Answer in Retrieval-Augmented Generation

TL;DR

CUE-R introduces an intervention-based framework to evaluate the operational utility of individual evidence items in retrieval-augmented generation, revealing their impact on answer correctness and faithfulness.

Contribution

The paper presents CUE-R, a novel lightweight method for measuring the utility of evidence items in RAG through targeted perturbations and utility axes analysis.

Findings

Removing or replacing evidence harms correctness and faithfulness.

Duplicating evidence often does not change behavior significantly.

Multi-hop evidence interactions are non-additive, affecting performance more than single supports.

Abstract

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.