When Do LLMs Need Retrieval Augmentation? Mitigating LLMs' Overconfidence Helps Retrieval Augmentation

TL;DR

This paper investigates when retrieval augmentation is necessary for LLMs by enhancing their awareness of knowledge limits, reducing overconfidence, and optimizing retrieval use for better efficiency and accuracy.

Contribution

It introduces methods to improve LLMs' perception of their knowledge boundaries, enabling more selective and efficient use of retrieval augmentation.

Findings

Enhanced LLMs better recognize their knowledge gaps.

Reduced reliance on retrieval without sacrificing performance.

Achieved comparable or improved results with fewer retrieval calls.

Abstract

Large Language Models (LLMs) have been found to have difficulty knowing they do not possess certain knowledge and tend to provide specious answers in such cases. Retrieval Augmentation (RA) has been extensively studied to mitigate LLMs' hallucinations. However, due to the extra overhead and unassured quality of retrieval, it may not be optimal to conduct RA all the time. A straightforward idea is to only conduct retrieval when LLMs are uncertain about a question. This motivates us to enhance the LLMs' ability to perceive their knowledge boundaries to help RA. In this paper, we first quantitatively measure LLMs' such ability and confirm their overconfidence. Then, we study how LLMs' certainty about a question correlates with their dependence on external retrieved information. We propose several methods to enhance LLMs' perception of knowledge boundaries and show that they are effective in reducing overconfidence. Additionally, equipped with these methods, LLMs can achieve comparable or even better performance of RA with much fewer retrieval calls.