Less LLM, More Documents: Searching for Improved RAG

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Ning, Jingjie, Kong, Yibo, Long, Yunfan, Callan, Jamie
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914514087182336
author Ning, Jingjie
Kong, Yibo
Long, Yunfan
Callan, Jamie
author_facet Ning, Jingjie
Kong, Yibo
Long, Yunfan
Callan, Jamie
contents Retrieval-Augmented Generation (RAG) couples document retrieval with large language models (LLMs). While scaling generators often improves accuracy, it also increases inference and deployment overhead. We study an orthogonal axis: enlarging the retriever's corpus, and how it trades off with generator scale. Across multiple open-domain QA benchmarks, corpus scaling consistently strengthens RAG and can in many cases match the gains of moving to a larger model tier, though with diminishing returns at larger scales. Small- and mid-sized generators paired with larger corpora often rival much larger models with smaller corpora; mid-sized models tend to gain the most, while tiny and very large models benefit less. Our analysis suggests that these improvements arise primarily from increased coverage of answer-bearing passages, while utilization efficiency remains largely unchanged. Overall, our results characterize a corpus-generator trade-off in RAG and provide empirical guidance on how corpus scale and model capacity interact in this setting.
format Preprint
id arxiv_https___arxiv_org_abs_2510_02657
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Less LLM, More Documents: Searching for Improved RAG
Ning, Jingjie
Kong, Yibo
Long, Yunfan
Callan, Jamie
Information Retrieval
Computation and Language
H.3.3; I.2.7
Retrieval-Augmented Generation (RAG) couples document retrieval with large language models (LLMs). While scaling generators often improves accuracy, it also increases inference and deployment overhead. We study an orthogonal axis: enlarging the retriever's corpus, and how it trades off with generator scale. Across multiple open-domain QA benchmarks, corpus scaling consistently strengthens RAG and can in many cases match the gains of moving to a larger model tier, though with diminishing returns at larger scales. Small- and mid-sized generators paired with larger corpora often rival much larger models with smaller corpora; mid-sized models tend to gain the most, while tiny and very large models benefit less. Our analysis suggests that these improvements arise primarily from increased coverage of answer-bearing passages, while utilization efficiency remains largely unchanged. Overall, our results characterize a corpus-generator trade-off in RAG and provide empirical guidance on how corpus scale and model capacity interact in this setting.
title Less LLM, More Documents: Searching for Improved RAG
topic Information Retrieval
Computation and Language
H.3.3; I.2.7
url https://arxiv.org/abs/2510.02657