OLAPH: Improving Factuality in Biomedical Long-form Question Answering

TL;DR

This paper introduces OLAPH, a novel framework for training large language models to improve factual accuracy in biomedical long-form question answering, using a new benchmark dataset and automatic evaluation methods.

Contribution

The paper presents OLAPH, a cost-effective, multifaceted evaluation framework that reduces hallucinations in LLMs and enhances factuality in biomedical long-form answers.

Findings

A 7B LLM trained with OLAPH matches medical experts' factuality in long answers.

OLAPH significantly improves LLM factuality on unseen evaluation metrics.

The MedLFQA dataset enables automatic evaluation of factual claims in biomedical QA.

Abstract

In the medical domain, numerous scenarios necessitate the long-form generation ability of large language models (LLMs). Specifically, when addressing patients' questions, it is essential that the model's response conveys factual claims, highlighting the need for an automated method to evaluate those claims. Thus, we introduce MedLFQA, a benchmark dataset reconstructed using long-form question-answering datasets related to the biomedical domain. We use MedLFQA to facilitate a cost-effective automatic evaluations of factuality. We also propose OLAPH, a simple and novel framework that utilizes cost-effective and multifaceted automatic evaluation to construct a synthetic preference set and answers questions in our preferred manner. Our framework leads us to train LLMs step-by-step to reduce hallucinations and include crucial medical claims. We highlight that, even on evaluation metrics not used during training, LLMs trained with our OLAPH framework demonstrate significant performance improvement in factuality. Our findings reveal that a 7B LLM trained with our OLAPH framework can provide long answers comparable to the medical experts' answers in terms of factuality. We believe that our work could shed light on gauging the long-text generation ability of LLMs in the medical domain. Our code and datasets are available.