Practical RAG Evaluation: A Rarity-Aware Set-Based Metric and Cost-Latency-Quality Trade-offs

TL;DR

This paper introduces a rarity-aware evaluation metric and a comprehensive benchmarking framework for production RAG systems, addressing limitations of classical IR metrics and providing reproducible, cost-aware decision tools.

Contribution

It proposes a new rarity-aware set score, a golden-set pipeline, and a detailed benchmark for production RAG, enabling better evaluation and optimization.

Findings

RA-nWG@K effectively measures rarity-aware retrieval quality.

The golden-set pipeline outperforms single-shot ranking methods.

Benchmark results reveal trade-offs among retrieval models, dimensions, and rerankers.

Abstract

This paper addresses the guessing game in building production RAG. Classical rank-centric IR metrics (nDCG/MAP/MRR) are a poor fit for RAG, where LLMs consume a set of passages rather than a browsed list; position discounts and prevalence-blind aggregation miss what matters: whether the prompt at cutoff K contains the decisive evidence. Second, there is no standardized, reproducible way to build and audit golden sets. Third, leaderboards exist but lack end-to-end, on-corpus benchmarking that reflects production trade-offs. Fourth, how state-of-the-art embedding models handle proper-name identity signals and conversational noise remains opaque. To address these, we contribute: (1) RA-nWG@K, a rarity-aware, per-query-normalized set score, and operational ceilings via the pool-restricted oracle ceiling (PROC) and the percentage of PROC (%PROC) to separate retrieval from ordering headroom within a Cost-Latency-Quality (CLQ) lens; (2) rag-gs (MIT), a lean golden-set pipeline with Plackett-Luce listwise refinement whose iterative updates outperform single-shot LLM ranking; (3) a comprehensive benchmark on a production RAG (scientific-papers corpus) spanning dense retrieval, hybrid dense+BM25, embedding models and dimensions, cross-encoder rerankers, ANN (HNSW), and quantization; and (4) targeted diagnostics that quantify proper-name identity signal and conversational-noise sensitivity via identity-destroying and formatting ablations. Together, these components provide practitioner Pareto guidance and auditable guardrails to support reproducible, budget/SLA-aware decisions.