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

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Feng, Naihe, Sui, Yi, Hou, Shiyi, Cresswell, Jesse C., Wu, Ga
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866909660306472960
author Feng, Naihe
Sui, Yi
Hou, Shiyi
Cresswell, Jesse C.
Wu, Ga
author_facet Feng, Naihe
Sui, Yi
Hou, Shiyi
Cresswell, Jesse C.
Wu, Ga
contents 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.
format Preprint
id arxiv_https___arxiv_org_abs_2506_20978
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Response Quality Assessment for Retrieval-Augmented Generation via Conditional Conformal Factuality
Feng, Naihe
Sui, Yi
Hou, Shiyi
Cresswell, Jesse C.
Wu, Ga
Information Retrieval
H.3.3
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.
title Response Quality Assessment for Retrieval-Augmented Generation via Conditional Conformal Factuality
topic Information Retrieval
H.3.3
url https://arxiv.org/abs/2506.20978