meditron-fo-anon/fully-open-meditron

Fully Open Meditron Corpus 是一个经过临床专家审核的训练语料库，专为医疗大型语言模型设计。它结合了八个公开的医疗问答数据集（包括MedQA、MedMCQA、PubMedQA、MedExpQA、HealthSearchQA、LiveQA、AfriMed-QA v1/v2）和三个临床专家审核的合成组件，总计约601,000个示例（约1.5亿个标记）。该语料库旨在支持临床决策支持和医疗问答的大型语言模型的监督微调，并确保数据来源、生成提示和去污染过程的完全透明性。它包含四个主要配置：curated_qa（聚合的公开医疗问答训练数据）、synthetic_curated_qa（基于公开数据生成的考试风格问答）、guidelines_qa（基于全球临床实践指南的问答）和synthetic_moove（开放式临床案例推理提示）。数据集采用OpenAI风格的对话格式，并提供了详细的元数据字段，如来源组件、合成标志和去污染信息。

A clinician-vetted training corpus for medical large language models, accompanying the paper *Fully Open Meditron: An Auditable Pipeline for Clinical LLMs* (anonymous submission to NeurIPS 2026 Evaluations & Datasets Track). The corpus combines eight aggregated public medical QA datasets with three clinician-vetted synthetic components, totaling approximately 601k examples (~150M tokens). It is designed to support supervised fine-tuning of large language models for clinical decision support and medical question answering, with full transparency over data provenance, generation prompts, and decontamination.