Uncertainty Quantification for In-Context Learning of Large Language Models

TL;DR

This paper introduces a novel method to quantify both aleatoric and epistemic uncertainties in in-context learning of large language models, enhancing trustworthiness and understanding of model responses.

Contribution

It proposes a new formulation and estimation technique for uncertainty quantification in in-context learning, addressing both data and model uncertainties in an unsupervised, plug-and-play manner.

Findings

Effective decomposition of uncertainties demonstrated through extensive experiments.

Unsupervised approach provides insights into LLM response reliability.

Method applicable without supervision or retraining.

Abstract

In-context learning has emerged as a groundbreaking ability of Large Language Models (LLMs) and revolutionized various fields by providing a few task-relevant demonstrations in the prompt. However, trustworthy issues with LLM's response, such as hallucination, have also been actively discussed. Existing works have been devoted to quantifying the uncertainty in LLM's response, but they often overlook the complex nature of LLMs and the uniqueness of in-context learning. In this work, we delve into the predictive uncertainty of LLMs associated with in-context learning, highlighting that such uncertainties may stem from both the provided demonstrations (aleatoric uncertainty) and ambiguities tied to the model's configurations (epistemic uncertainty). We propose a novel formulation and corresponding estimation method to quantify both types of uncertainties. The proposed method offers an unsupervised way to understand the prediction of in-context learning in a plug-and-play fashion. Extensive experiments are conducted to demonstrate the effectiveness of the decomposition. The code and data are available at: https://github.com/lingchen0331/UQ_ICL.