Understanding the Impact of Confidence in Retrieval Augmented Generation: A Case Study in the Medical Domain

Accepted at BioNLP 2025 (Workshop cohosted with ACL2025)
Shintaro Ozaki1, Yuta Kato2, Siyuan Feng2, Masayo Tomita2, Kazuki Hayashi1, Wataru Hashimoto1,
1Nara Institute of Science and Technology 2The University of Tokyo 3NEC Corporation
Teaser image

The focus of our research is to analyze whether RAG improves the confidence of the model response.

Abstract

Retrieval Augmented Generation (RAG) complements the knowledge of Large Language Models (LLMs) by leveraging external information to enhance response accuracy for queries.
This approach is widely applied in several fields by taking its advantage of injecting the most up-to-date information, and researchers are focusing on understanding and improving this aspect to unlock the full potential of RAG in such high-stakes applications.
However, despite the potential of RAG to address these needs, the mechanisms behind the confidence levels of its outputs remain underexplored.
Our study focuses on the impact of RAG, specifically examining whether RAG improves the confidence of LLM outputs in the medical domain.
We conduct this analysis across various configurations and models.
We evaluate confidence by treating the model's predicted probability as its output and calculating several evaluation metrics which include calibration error method, entropy, the best probability, and accuracy.
Experimental results across multiple datasets confirmed that certain models possess the capability to judge for themselves whether an inserted document relates to the correct answer.
These results suggest that evaluating models based on their output probabilities determine whether they function as generators in the RAG framework.
Our approach allows us to evaluate whether the models handle retrieved documents.

BibTeX

@misc{ozaki2024understandingimpactconfidenceretrieval,
        title={Understanding the Impact of Confidence in Retrieval Augmented Generation: A Case Study in the Medical Domain}, 
        author={Shintaro Ozaki and Yuta Kato and Siyuan Feng and Masayo Tomita and Kazuki Hayashi and Ryoma Obara and Masafumi Oyamada and Katsuhiko Hayashi and Hidetaka Kamigaito and Taro Watanabe},
        year={2024},
        eprint={2412.20309},
        archivePrefix={arXiv},
        primaryClass={cs.CL},
        url={https://arxiv.org/abs/2412.20309}, 
  }