The Aloe Family Recipe for Open and Specialized Healthcare LLMs

Purpose: With advancements in Large Language Models (LLMs) for healthcare,the need arises for competitive open-source models to protect the publicinterest. This work contributes to the field of open medical LLMs by optimizingkey stages of data preprocessing and training, while showing how to improvemodel safety (through DPO) and efficacy (through RAG). The evaluationmethodology used, which includes four different types of tests, defines a newstandard for the field. The resultant models, shown to be competitive with thebest private alternatives, are released with a permisive license. Methods: Building on top of strong base models like Llama 3.1 and Qwen 2.5,Aloe Beta uses a custom dataset to enhance public data with synthetic Chain ofThought examples. The models undergo alignment with Direct PreferenceOptimization, emphasizing ethical and policy-aligned performance in thepresence of jailbreaking attacks. Evaluation includes close-ended, open-ended,safety and human assessments, to maximize the reliability of results. Results: Recommendations are made across the entire pipeline, backed by thesolid performance of the Aloe Family. These models deliver competitiveperformance across healthcare benchmarks and medical fields, and are oftenpreferred by healthcare professionals. On bias and toxicity, the Aloe Betamodels significantly improve safety, showing resilience to unseen jailbreakingattacks. For a responsible release, a detailed risk assessment specific tohealthcare is attached to the Aloe Family models. Conclusion: The Aloe Beta models, and the recipe that leads to them, are asignificant contribution to the open-source medical LLM field, offeringtop-of-the-line performance while maintaining high ethical requirements. Thiswork sets a new standard for developing and reporting aligned LLMs inhealthcare.