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.

Discriminating the Concepts of Knowledge Retrieval

Research on Methodology of Fuzzy Retrieve Based on Soft Computing

Ontology-based Related Knowledge Visualization Retrieval Model

A Review of Knowledge Transfer Based on Foreign Studies

Knowledge and Knowledge Management: a Literature Review

Ontology-based Product Knowledge Representation and Retrieval

Applying Information Retrieval Technology To Incremental Knowledge Management

Research on the Knowledge Management in Component Retrieval by Intelligent Agent

Enhancing database management to knowledge base management: The role of information retrieval technology

Research on Knowledge Representation and Knowledge Relevance Measurement

Discriminative Information Retrieval for Knowledge Discovery

An Empirical Investigation of Implicit and Explicit Knowledge-Enhanced Methods for Ad Hoc Dataset Retrieval

Information Retrieval in the Grid Environment

Knowledge Ontology Establishment of Collaborative and Application

Some Information Retrieval Terms

STATE OF THE ART OF INFORMATION RETRIEVAL ON THE WEB

A Review of Advanced Topics in Information Retrieval

The Intercrossing of Disciplines & Automatic Information Retrieval

An Intelligent Platform For Information Retrieval

Ontology-based construction knowledge retrieval system

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