Pandora's Box or Aladdin's Lamp: A Comprehensive Analysis Revealing the Role of RAG Noise in Large Language Models

TL;DR

This paper introduces a comprehensive evaluation framework called NoiserBench for analyzing the impact of seven distinct noise types on large language models in retrieval-augmented generation, revealing that some noise can be beneficial while others are harmful.

Contribution

It defines seven noise types from a linguistic perspective, establishes a new benchmark, and empirically evaluates their effects on diverse LLMs, uncovering that some noise can improve model performance.

Findings

Beneficial noise can enhance LLM capabilities.

Harmful noise generally impairs performance.

Some noise types are actually advantageous for LLMs.

Abstract

Retrieval-Augmented Generation (RAG) has emerged as a crucial method for addressing hallucinations in large language models (LLMs). While recent research has extended RAG models to complex noisy scenarios, these explorations often confine themselves to limited noise types and presuppose that noise is inherently detrimental to LLMs, potentially deviating from real-world retrieval environments and restricting practical applicability. In this paper, we define seven distinct noise types from a linguistic perspective and establish a Noise RAG Benchmark (NoiserBench), a comprehensive evaluation framework encompassing multiple datasets and reasoning tasks. Through empirical evaluation of eight representative LLMs with diverse architectures and scales, we reveal that these noises can be further categorized into two practical groups: noise that is beneficial to LLMs (aka beneficial noise) and noise that is harmful to LLMs (aka harmful noise). While harmful noise generally impairs performance, beneficial noise may enhance several aspects of model capabilities and overall performance. Our analysis offers insights for developing more robust, adaptable RAG solutions and mitigating hallucinations across diverse retrieval scenarios. Code is available at https://github.com/jinyangwu/NoiserBench.