Multi-LLM Adaptive Conformal Inference for Reliable LLM Responses

TL;DR

This paper introduces MACI, a new conformal inference method that improves the reliability and efficiency of factuality assessments in large language models, especially in high-stakes domains.

Contribution

We reformulate conformal inference as a multiplicative filtering problem and develop MACI, which uses ensembles for better factuality scoring and group-conditional calibration for validity.

Findings

MACI achieves higher retention rates than baselines.

MACI maintains validity with user-specified coverage.

MACI reduces time cost compared to existing methods.

Abstract

Ensuring factuality is essential for the safe use of Large Language Models (LLMs) in high-stakes domains such as medicine and law. Conformal inference provides distribution-free guarantees, but existing approaches are either overly conservative, discarding many true-claims, or rely on adaptive error rates and simple linear models that fail to capture complex group structures. To address these challenges, we reformulate conformal inference in a multiplicative filtering setting, modeling factuality as a product of claim-level scores. Our method, Multi-LLM Adaptive Conformal Inference (MACI), leverages ensembles to produce more accurate factuality-scores, which in our experiments led to higher retention, while validity is preserved through group-conditional calibration. Experiments show that MACI consistently achieves user-specified coverage with substantially higher retention and lower time cost than baselines. Our repository is available at https://github.com/MLAI-Yonsei/MACI