# Coverage, Not Averages: Semantic Stratification for Trustworthy Retrieval Evaluation

arXiv: https://arxiv.org/abs/2604.20763 · HTML: https://mlsanity.com/paper/2604.20763/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in RAG robustness and adaptive retrieval, one of 15 directions in Retrieval-augmented generation.
- Builds on 18 of its 33 references that are in the map.
- First posted on arXiv 22 Apr 2026.

## Placed in

- Retrieval-augmented generation / [RAG robustness and adaptive retrieval](https://mlsanity.com/rag/rag-robustness-and-adaptive-retrieval/)

## Builds on (most cited first, 18 of 18)

- [HotpotQA: A Dataset for Diverse, Explainable Multi-hop Question Answering](https://mlsanity.com/paper/1809.09600/) (2018)
- [TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehension](https://mlsanity.com/paper/1705.03551/) (2017)
- [Billion-scale similarity search with GPUs](https://mlsanity.com/paper/1702.08734/) (2017)
- [MS MARCO: A Human Generated MAchine Reading COmprehension Dataset](https://mlsanity.com/paper/1611.09268/) (2016)
- [From Local to Global: A Graph RAG Approach to Query-Focused Summarization](https://mlsanity.com/paper/2404.16130/) (2024)
- [BEIR: A Heterogenous Benchmark for Zero-shot Evaluation of Information Retrieval Models](https://mlsanity.com/paper/2104.08663/) (2021)
- [SPECTER: Document-level Representation Learning using Citation-informed Transformers](https://mlsanity.com/paper/2004.07180/) (2020)
- [From Louvain to Leiden: guaranteeing well-connected communities](https://mlsanity.com/paper/1810.08473/) (2018)
- [tinyBenchmarks: evaluating LLMs with fewer examples](https://mlsanity.com/paper/2402.14992/) (2024)
- [HippoRAG: Neurobiologically Inspired Long-Term Memory for Large Language Models](https://mlsanity.com/paper/2405.14831/) (2024)
- [MultiHop-RAG: Benchmarking Retrieval-Augmented Generation for Multi-Hop Queries](https://mlsanity.com/paper/2401.15391/) (2024)
- [CRUD-RAG: A Comprehensive Chinese Benchmark for Retrieval-Augmented Generation of Large Language Models](https://mlsanity.com/paper/2401.17043/) (2024)
- [Long2RAG: Evaluating Long-Context & Long-Form Retrieval-Augmented Generation with Key Point Recall](https://mlsanity.com/paper/2410.23000/) (2024)
- [QE-RAG: A Robust Retrieval-Augmented Generation Benchmark for Query Entry Errors](https://mlsanity.com/paper/2504.04062/) (2025)
- [Know Your RAG: Dataset Taxonomy and Generation Strategies for Evaluating RAG Systems](https://mlsanity.com/paper/2411.19710/) (2024)
- [Towards Global Retrieval Augmented Generation: A Benchmark for Corpus-Level Reasoning](https://mlsanity.com/paper/2510.26205/) (2025)
- [Efficient Test Collection Construction via Active Learning](https://mlsanity.com/paper/1801.05605/) (2018)
- [mmRAG: A Modular Benchmark for Retrieval-Augmented Generation over Text, Tables, and Knowledge Graphs](https://mlsanity.com/paper/2505.11180/) (2025)

## Built on by (most cited first, 2 of 2)

- [Where Does Retrieval Fail? Evaluating RAG Architectures for Agricultural Advisory](https://mlsanity.com/paper/2608.14886/) (2026)
- [CoverSynth: Coverage-Guided Generation of Retrieval Evaluation Sets](https://mlsanity.com/paper/2610.15409/) (2026)

## Abstract (arXiv)

Retrieval quality is the primary bottleneck for accuracy and robustness in retrieval-augmented generation (RAG). Current evaluation relies on heuristically constructed query sets, which introduce a hidden intrinsic bias. We formalize retrieval evaluation as a statistical estimation problem, showing that metric reliability is fundamentally limited by the evaluation-set construction. We further introduce semantic stratification, which grounds evaluation in corpus structure by organizing documents into an interpretable global space of entity-based clusters and systematically generating queries for missing strata. This yields (1) formal semantic coverage guarantees across retrieval regimes and (2) interpretable visibility into retrieval failure modes. Experiments across multiple benchmarks and retrieval methods validate our framework. The results expose systematic coverage gaps, identify structural signals that explain variance in retrieval performance, and show that stratified evaluation yields more stable and transparent assessments while supporting more trustworthy decision-making than aggregate metrics.