Evaluating the Retrieval Robustness of Large Language Models

2025-05-28

AI Summary

This work systematically evaluates the robustness of large language models (LLMs) in retrieval-augmented generation (RAG), addressing three critical questions: whether RAG consistently outperforms non-RAG baselines, whether increasing retrieved document count always improves performance, and whether document ordering affects generation quality. Method: We introduce the novel concept of *retrieval robustness*, construct a benchmark dataset of 1,500 open-domain questions, and propose three dedicated quantitative metrics. Using Wikipedia-based retrieval, we conduct large-scale controlled experiments across 11 LLMs and 3 prompting strategies. Contribution/Results: All models exhibit non-negligible retrieval robustness, yet remain sensitive to retrieval noise and document structure—severely limiting RAG’s practical gains. Our findings expose fundamental deployment bottlenecks in real-world RAG systems and provide both theoretical insights and empirical evidence for designing robust RAG architectures.

Technology Category

• Natural Language Processing: Safety and Robustness
• Machine Learning: Large Multimodal Models (LMMs)
• Data Mining & Knowledge Management: Conversational Systems for Recommendation & Retrieval

Application Category

• Search and Retrieval-Augmented AI: Retrieval-Augmented Generation (RAG) and multi-modal RAG
• Web Mining and Content Analysis: Robustness and generalizability of Web mining methods
• User Modeling, Personalization and Recommendation: Fairness-aware retrieval and ranking

Abstract

Retrieval-augmented generation (RAG) generally enhances large language models' (LLMs) ability to solve knowledge-intensive tasks. But RAG may also lead to performance degradation due to imperfect retrieval and the model's limited ability to leverage retrieved content. In this work, we evaluate the robustness of LLMs in practical RAG setups (henceforth retrieval robustness). We focus on three research questions: (1) whether RAG is always better than non-RAG; (2) whether more retrieved documents always lead to better performance; (3) and whether document orders impact results. To facilitate this study, we establish a benchmark of 1500 open-domain questions, each with retrieved documents from Wikipedia. We introduce three robustness metrics, each corresponds to one research question. Our comprehensive experiments, involving 11 LLMs and 3 prompting strategies, reveal that all of these LLMs exhibit surprisingly high retrieval robustness; nonetheless, different degrees of imperfect robustness hinders them from fully utilizing the benefits of RAG.

Problem

• Assessing if RAG consistently outperforms non-RAG approaches
• Examining if more retrieved documents guarantee better performance
• Investigating how document order affects retrieval-augmented results

Innovation

• Evaluates LLM robustness in RAG setups
• Introduces three retrieval robustness metrics
• Tests 11 LLMs with 3 prompting strategies