Pareto Optimal Learning for Estimating Large Language Model Errors

TL;DR

This paper introduces a Pareto optimization-based method to estimate the error probability of Large Language Model responses, improving error detection and correction by integrating multiple information sources.

Contribution

It presents a novel Pareto optimal framework for error estimation in LLMs, aligning multiple information sources with the model's errors, and enhances LLM performance through dynamic risk scoring.

Findings

Risk scores correlate well with true LLM errors

Method surpasses state-of-the-art task-specific models

Enables effective error correction and performance improvement

Abstract

Large Language Models (LLMs) have shown impressive abilities in many applications. When a concrete and precise answer is desired, it is important to have a quantitative estimation of the potential error rate. However, this can be challenging due to the text-in-text-out nature of generative models. We present a method based on Pareto optimization that generates a risk score to estimate the probability of error in an LLM response by integrating multiple sources of information. We prove theoretically that the error estimator optimized in our framework aligns with the LLM and the information sources in an Pareto optimal manner. Experimental results show that the risk scores estimated by our method are well correlated with the true LLM error rate, thus facilitating error correction. By dynamically combining with prompting strategies such as self-verification and information retrieval, we demonstrate the proposed method can be utilized to increase the performance of an LLM, surpassing state-of-the-art task specific models.