How do you know that? Teaching Generative Language Models to Reference Answers to Biomedical Questions

TL;DR

This paper presents a retrieval-augmented generation system for biomedical question answering that improves answer reliability by referencing relevant PubMed abstracts, enabling verification and enhancing factual accuracy.

Contribution

It introduces a fine-tuned LLM system that references PubMed abstracts in answers, improving reliability and transparency in biomedical question answering.

Findings

Retrieval system improves PubMed search accuracy by 23%.

Fine-tuned LLM achieves comparable referencing to GPT-4 Turbo.

Publicly available dataset and models for biomedical QA.

Abstract

Large language models (LLMs) have recently become the leading source of answers for users' questions online. Despite their ability to offer eloquent answers, their accuracy and reliability can pose a significant challenge. This is especially true for sensitive domains such as biomedicine, where there is a higher need for factually correct answers. This paper introduces a biomedical retrieval-augmented generation (RAG) system designed to enhance the reliability of generated responses. The system is based on a fine-tuned LLM for the referenced question-answering, where retrieved relevant abstracts from PubMed are passed to LLM's context as input through a prompt. Its output is an answer based on PubMed abstracts, where each statement is referenced accordingly, allowing the users to verify the answer. Our retrieval system achieves an absolute improvement of 23% compared to the PubMed search engine. Based on the manual evaluation on a small sample, our fine-tuned LLM component achieves comparable results to GPT-4 Turbo in referencing relevant abstracts. We make the dataset used to fine-tune the models and the fine-tuned models based on Mistral-7B-instruct-v0.1 and v0.2 publicly available.