exafluence 's Collections

US Med-AI Models and Datasets

A collection of fine-tuned LLMs and datasets for US medical applications, enhancing tasks like medical coding, clinical support, and medical QA.


  • Note Open Nexus MedQA is a comprehensive dataset curated for medical question-answering models. It integrates data from various public medical datasets such as ChatDoctor, icliniq, CareQA, and others, providing a diverse set of medical queries, instructions, and accurate responses. Designed specifically to advance AI in the healthcare domain, Open Nexus MedQA enables training and evaluation of large language models (LLMs) for tasks like medical diagnosis, patient support, and medical coding.


  • Note EXF Medistral Nemo 12B is a powerful large language model based on the Mistral-Nemo 12B architecture, fine-tuned specifically for applications in the medical domain. It is trained on datasets such as Open-Nexus-MedQA, which integrates a wide range of medical queries and knowledge sources. This model is designed to assist in tasks like medical question answering, clinical decision support, and medical coding.


  • Note OASISCoder is a fine-tuned language model based on the LLaMA 3.2 3B architecture, optimized for handling CMS (Centers for Medicare & Medicaid Services) and OASIS medical coding tasks. It is designed to assist with medical documentation, question-answering, and regulatory compliance in the healthcare domain, making it useful for automating or improving the accuracy of medical coding processes.