Saved in:
Bibliographic Details
Main Authors: Bennani, Sofia, Moslonka, Charles
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2601.14123
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909995856035840
author Bennani, Sofia
Moslonka, Charles
author_facet Bennani, Sofia
Moslonka, Charles
contents We study how document chunking choices impact the reliability of Retrieval-Augmented Generation (RAG) systems in industry. While practice often relies on heuristics, our end-to-end evaluation on Natural Questions systematically varies chunking method (token, sentence, semantic, code), chunk size, overlap, and context length. We use a standard industrial setup: SPLADE retrieval and a Mistral-8B generator. We derive actionable lessons for cost-efficient deployment: (i) overlap provides no measurable benefit and increases indexing cost; (ii) sentence chunking is the most cost-effective method, matching semantic chunking up to ~5k tokens; (iii) a "context cliff" reduces quality beyond ~2.5k tokens; and (iv) optimal context depends on the goal (semantic quality peaks at small contexts; exact match at larger ones).
format Preprint
id arxiv_https___arxiv_org_abs_2601_14123
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle A Systematic Analysis of Chunking Strategies for Reliable Question Answering
Bennani, Sofia
Moslonka, Charles
Computation and Language
Information Retrieval
We study how document chunking choices impact the reliability of Retrieval-Augmented Generation (RAG) systems in industry. While practice often relies on heuristics, our end-to-end evaluation on Natural Questions systematically varies chunking method (token, sentence, semantic, code), chunk size, overlap, and context length. We use a standard industrial setup: SPLADE retrieval and a Mistral-8B generator. We derive actionable lessons for cost-efficient deployment: (i) overlap provides no measurable benefit and increases indexing cost; (ii) sentence chunking is the most cost-effective method, matching semantic chunking up to ~5k tokens; (iii) a "context cliff" reduces quality beyond ~2.5k tokens; and (iv) optimal context depends on the goal (semantic quality peaks at small contexts; exact match at larger ones).
title A Systematic Analysis of Chunking Strategies for Reliable Question Answering
topic Computation and Language
Information Retrieval
url https://arxiv.org/abs/2601.14123