Tug-of-War Between Knowledge: Exploring and Resolving Knowledge Conflicts in Retrieval-Augmented Language Models

TL;DR

This paper investigates how retrieval-augmented language models handle conflicting knowledge from internal memory and external sources, revealing biases and proposing a method to resolve these conflicts effectively.

Contribution

It introduces an evaluation framework for knowledge conflicts in RALMs, analyzes their behavior and biases, and proposes the Conflict-Disentangle Contrastive Decoding (CD2) method to resolve conflicts.

Findings

RALMs favor internal memory even with correct external evidence

RALMs exhibit availability and confirmation biases

CD2 effectively resolves knowledge conflicts

Abstract

Retrieval-augmented language models (RALMs) have demonstrated significant potential in refining and expanding their internal memory by retrieving evidence from external sources. However, RALMs will inevitably encounter knowledge conflicts when integrating their internal memory with external sources. Knowledge conflicts can ensnare RALMs in a tug-of-war between knowledge, limiting their practical applicability. In this paper, we focus on exploring and resolving knowledge conflicts in RALMs. First, we present an evaluation framework for assessing knowledge conflicts across various dimensions. Then, we investigate the behavior and preference of RALMs from the following two perspectives: (1) Conflicts between internal memory and external sources: We find that stronger RALMs emerge with the Dunning-Kruger effect, persistently favoring their faulty internal memory even when correct evidence is provided. Besides, RALMs exhibit an availability bias towards common knowledge; (2) Conflicts between truthful, irrelevant and misleading evidence: We reveal that RALMs follow the principle of majority rule, leaning towards placing trust in evidence that appears more frequently. Moreover, we find that RALMs exhibit confirmation bias, and are more willing to choose evidence that is consistent with their internal memory. To solve the challenge of knowledge conflicts, we propose a method called Conflict-Disentangle Contrastive Decoding (CD2) to better calibrate the model's confidence. Experimental results demonstrate that our CD2 can effectively resolve knowledge conflicts in RALMs.