Multi-LLM Adaptive Conformal Inference for Reliable LLM Responses
This addresses the need for reliable LLM responses in critical fields like medicine and law, representing an incremental improvement over existing conformal inference methods.
The paper tackles the problem of ensuring factuality in Large Language Models for high-stakes domains by proposing Multi-LLM Adaptive Conformal Inference (MACI), which reformulates conformal inference to model factuality as a product of claim-level scores, resulting in higher retention and lower time cost while preserving validity.
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