Benchmarking large language models for biomedical natural language processing applications and recommendations

TL;DR

This paper systematically evaluates large language models in biomedical NLP tasks, comparing their performance with traditional models, and provides practical insights and recommendations for their application in the biomedical domain.

Contribution

It offers a comprehensive benchmark of LLMs in BioNLP, highlighting their strengths, limitations, and the importance of fine-tuning for optimal performance.

Findings

Traditional fine-tuning outperforms zero or few-shot LLMs in most tasks.

GPT-4 excels in reasoning-related biomedical tasks.

LLMs exhibit issues like hallucinations and missing information.

Abstract

The rapid growth of biomedical literature poses challenges for manual knowledge curation and synthesis. Biomedical Natural Language Processing (BioNLP) automates the process. While Large Language Models (LLMs) have shown promise in general domains, their effectiveness in BioNLP tasks remains unclear due to limited benchmarks and practical guidelines. We perform a systematic evaluation of four LLMs, GPT and LLaMA representatives on 12 BioNLP benchmarks across six applications. We compare their zero-shot, few-shot, and fine-tuning performance with traditional fine-tuning of BERT or BART models. We examine inconsistencies, missing information, hallucinations, and perform cost analysis. Here we show that traditional fine-tuning outperforms zero or few shot LLMs in most tasks. However, closed-source LLMs like GPT-4 excel in reasoning-related tasks such as medical question answering. Open source LLMs still require fine-tuning to close performance gaps. We find issues like missing information and hallucinations in LLM outputs. These results offer practical insights for applying LLMs in BioNLP.