RAG in the Wild: On the (In)effectiveness of LLMs with Mixture-of-Knowledge Retrieval Augmentation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Xu, Ran, Zhuang, Yuchen, Yu, Yue, Wang, Haoyu, Shi, Wenqi, Yang, Carl
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866918105279627264
author Xu, Ran
Zhuang, Yuchen
Yu, Yue
Wang, Haoyu
Shi, Wenqi
Yang, Carl
author_facet Xu, Ran
Zhuang, Yuchen
Yu, Yue
Wang, Haoyu
Shi, Wenqi
Yang, Carl
contents Retrieval-augmented generation (RAG) enhances large language models (LLMs) by integrating external knowledge retrieved at inference time. While RAG demonstrates strong performance on benchmarks largely derived from general-domain corpora like Wikipedia, its effectiveness under realistic, diverse retrieval scenarios remains underexplored. We evaluated RAG systems using MassiveDS, a large-scale datastore with mixture of knowledge, and identified critical limitations: retrieval mainly benefits smaller models, rerankers add minimal value, and no single retrieval source consistently excels. Moreover, current LLMs struggle to route queries across heterogeneous knowledge sources. These findings highlight the need for adaptive retrieval strategies before deploying RAG in real-world settings. Our code and data can be found at https://github.com/ritaranx/RAG_in_the_Wild.
format Preprint
id arxiv_https___arxiv_org_abs_2507_20059
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle RAG in the Wild: On the (In)effectiveness of LLMs with Mixture-of-Knowledge Retrieval Augmentation
Xu, Ran
Zhuang, Yuchen
Yu, Yue
Wang, Haoyu
Shi, Wenqi
Yang, Carl
Computation and Language
Artificial Intelligence
Machine Learning
Retrieval-augmented generation (RAG) enhances large language models (LLMs) by integrating external knowledge retrieved at inference time. While RAG demonstrates strong performance on benchmarks largely derived from general-domain corpora like Wikipedia, its effectiveness under realistic, diverse retrieval scenarios remains underexplored. We evaluated RAG systems using MassiveDS, a large-scale datastore with mixture of knowledge, and identified critical limitations: retrieval mainly benefits smaller models, rerankers add minimal value, and no single retrieval source consistently excels. Moreover, current LLMs struggle to route queries across heterogeneous knowledge sources. These findings highlight the need for adaptive retrieval strategies before deploying RAG in real-world settings. Our code and data can be found at https://github.com/ritaranx/RAG_in_the_Wild.
title RAG in the Wild: On the (In)effectiveness of LLMs with Mixture-of-Knowledge Retrieval Augmentation
topic Computation and Language
Artificial Intelligence
Machine Learning
url https://arxiv.org/abs/2507.20059