AlpaCare:Instruction-tuned Large Language Models for Medical Application
This addresses the problem of poor generalizability in medical LLMs for healthcare applications, though it is incremental as it builds on existing IFT methods with a new dataset.
The paper tackled the limited diversity in medical instruction-finetuning datasets by creating MedInstruct-52k using GPT-4 and ChatGPT, and fine-tuning LLaMA models to develop AlpaCare, which achieved up to 38.1% absolute gain in medical evaluations and 6.7% average gain on general benchmarks.
Instruction-finetuning (IFT) has become crucial in aligning Large Language Models (LLMs) with diverse human needs and has shown great potential in medical applications. However, previous studies mainly fine-tune LLMs on biomedical datasets with limited diversity, which often rely on benchmarks or narrow task scopes, and hence significantly limit the effectiveness on their medical instruction-following ability and generalizability. To bridge this gap, we propose creating a diverse, machine-generated medical IFT dataset, MedInstruct-52k, using GPT-4 and ChatGPT with a high-quality expert-curated seed set. We then fine-tune LLaMA-series models on the dataset to develop AlpaCare. Despite using a smaller domain-specific dataset than previous medical LLMs, AlpaCare not only demonstrates superior performance on medical applications, with up to 38.1% absolute gain over best baselines in medical free-form instruction evaluations, but also achieves 6.7% absolute gains averaged over multiple general domain benchmarks. Human evaluation further shows that AlpaCare consistently outperforms best baselines in terms of both correctness and helpfulness. We offer public access to our data, model, and codebase in https://github.com/XZhang97666/AlpaCare.