Response Quality Assessment for Retrieval-Augmented Generation via Conditional Conformal Factuality

TL;DR

This paper introduces Conformal-RAG, a new framework that applies conformal prediction to Retrieval-Augmented Generation, providing statistical guarantees on response quality and sub-claim reliability without needing ground truth answers.

Contribution

Conformal-RAG is the first method to incorporate conformal prediction into RAG, ensuring probabilistic guarantees on sub-claim quality without manual labeling or ground truth.

Findings

Conformal-RAG retains up to 60% more high-quality sub-claims.

It provides statistical guarantees on response reliability.

It does not require manual labeling of conformal sets.

Abstract

Existing research on Retrieval-Augmented Generation (RAG) primarily focuses on improving overall question-answering accuracy, often overlooking the quality of sub-claims within generated responses. Recent methods that attempt to improve RAG trustworthiness, such as through auto-evaluation metrics, lack probabilistic guarantees or require ground truth answers. To address these limitations, we propose Conformal-RAG, a novel framework inspired by recent applications of conformal prediction (CP) on large language models (LLMs). Conformal-RAG leverages CP and internal information from the RAG mechanism to offer statistical guarantees on response quality. It ensures group-conditional coverage spanning multiple sub-domains without requiring manual labelling of conformal sets, making it suitable for complex RAG applications. Compared to existing RAG auto-evaluation methods, Conformal-RAG offers statistical guarantees on the quality of refined sub-claims, ensuring response reliability without the need for ground truth answers. Additionally, our experiments demonstrate that by leveraging information from the RAG system, Conformal-RAG retains up to 60\% more high-quality sub-claims from the response compared to direct applications of CP to LLMs, while maintaining the same reliability guarantee.