Kanzoet97/Melon

Medical-Reasoning-SFT-GPT-OSS-120B 是一个高质量的合成数据集，包含使用 OpenAI 的 gpt-oss-120B 模型生成的医疗推理对话，推理努力设置为 high，专为医疗保健应用中大型语言模型的监督微调而设计。数据集涵盖了广泛的医学领域，包括临床医学、基础科学、诊断学、医学教育和研究。每个样本遵循标准聊天格式，展示了结构化的医学思维和逐步推理过程。数据集统计显示总样本数为 200,927，总标记数为 539,165,577，平均每个样本的标记数为 2,683.3。

Medical-Reasoning-SFT-GPT-OSS-120B is a high-quality synthetic dataset of medical reasoning conversations generated using OpenAIs gpt-oss-120B model with reasoning effort set to high, designed for supervised fine-tuning of large language models in healthcare applications. The dataset covers a wide range of medical domains including clinical medicine, basic sciences, diagnostics, medical education, and research. Each sample follows a standard chat format, demonstrating structured medical thinking with step-by-step reasoning processes. Dataset statistics show a total of 200,927 samples, 539,165,577 tokens, and an average of 2,683.3 tokens per sample.