Evaluating the Utility of Grounding Documents with Reference-Free LLM-based Metrics
This addresses the challenge of evaluating grounding documents in RAG systems for AI practitioners, offering an incremental improvement over existing metrics.
The paper tackles the problem of quantifying content utility in Retrieval Augmented Generation (RAG) by proposing GroGU, a reference-free metric based on LLM generation confidence, which improves query-rewriting for RAG with gains of up to 18.2 points in Mean Reciprocal Rank and 9.4 points in answer accuracy.
Retrieval Augmented Generation (RAG)'s success depends on the utility the LLM derives from the content used for grounding. Quantifying content utility does not have a definitive specification and existing metrics ignore model-specific capabilities and/or rely on costly annotations. In this paper, we propose Grounding Generation Utility (GroGU), a model-specific and reference-free metric that defines utility as a function of the downstream LLM's generation confidence based on entropy. Despite having no annotation requirements, GroGU is largely faithful in distinguishing ground-truth documents while capturing nuances ignored by LLM-agnostic metrics. We apply GroGU to train a query-rewriter for RAG by identifying high-utility preference data for Direct Preference Optimization. Experiments show improvements by up to 18.2 points in Mean Reciprocal Rank and up to 9.4 points in answer accuracy.