# Study finds retrieval, not reading, limits RAG in long-context QA

Digest AI · Research · published 2026-10-21T04:00:00Z

Canonical: https://digestai.news/story/study-finds-retrieval-not-reading-limits-rag-in-long-context-qa

## Summary

Researchers introduced evidence-based generation (EBG) to diagnose why retrieval-augmented generation (RAG) struggles with long-context question answering. The study argues that the bottleneck is not the model's ability to read long texts, but the RAG pipeline's inability to retrieve all necessary evidence chunks. By requiring models to provide line-level citations, the team created a joint Ans+Evid metric that scores both answer correctness and grounding.

The team reformulated existing corpora into new benchmarks, including MultiHop-RAG and QASPER-MP, and added synthetic datasets ranging from 10K to 1M characters. They tested ten proprietary frontier models and three open-weight large language models. The results show that similarity-based retrieval often fails to surface all required chunks for multi-fact queries, whereas long-context models can handle these tasks without chunking. Additionally, chunking destroys document-level meta-information needed to resolve references across passages.

The study highlights that standard answer-only evaluations miss significant grounding failures, such as incomplete citations in complex settings. When context contradicts a model's internal knowledge, the gap between simple answer accuracy and grounded accuracy widens. The authors have released the data and code for this work on GitHub to support further research into evidence-grounded evaluation.

## Key points

- EBG metric scores answer correctness and grounding simultaneously using line-level citations.
- Similarity retrieval fails to surface all required chunks for multi-fact queries in RAG.
- Chunking destroys document meta-information needed to resolve cross-passage references.

## Why it matters

This research identifies specific structural flaws in RAG pipelines, suggesting that long-context models may outperform chunked retrieval for complex, multi-document tasks. It provides a new evaluation framework for assessing how well models ground their answers in source material.

## Sources

1. [The Bottleneck Is Retrieval, Not Reading: Diagnosing Evidence-Based Generation with RAG for Long Context Question Answering](https://arxiv.org/abs/2610.26951) (arXiv cs.CL, 2026-10-21, primary source)

## Cite

Digest AI, "Study finds retrieval, not reading, limits RAG in long-context QA", 21 October 2026, https://digestai.news/story/study-finds-retrieval-not-reading-limits-rag-in-long-context-qa

---

Written by Digest AI's editorial model from the linked sources; the sources are the record. Headlines, digests and key points are written by Digest AI and may be quoted with a link to the story page. Linked articles belong to their publishers. Terms: https://digestai.news/terms#reuse
JSON: https://digestai.news/story/study-finds-retrieval-not-reading-limits-rag-in-long-context-qa.json