MMedS-Bench

💻Github Repo 🖨️arXiv Paper

The official instruction finetuning dataset for "Towards Evaluating and Building Versatile Large Language Models for Medicine".

Introduction

The MedS-Ins dataset is a meticulously curated instruction-tuning dataset, specifically crafted to enhance the capabilities of large language models (LLMs) in handling complex medical tasks. This dataset draws from a diverse array of text domains, encompassing exams, clinical texts, academic papers, medical knowledge bases, and daily conversations. These domains have been carefully selected to represent a wide spectrum of medical knowledge and interaction, providing a well-rounded foundation for training medical LLMs.

Data Format

The data format is the same as MedS-Ins.

{
  "Contributors": [""],
  "Source": [""],
  "URL": [""],
  "Categories": [""],
  "Reasoning": [""],
  "Definition": [""],
  "Input_language": [""], 
  "Output_language": [""],
  "Instruction_language": [""],  
  "Domains": [""],    
  "Positive Examples": [ { "input": "", "output": "",  "explanation": ""} ], 
  "Negative Examples": [ { "input": "", "output": "",  "explanation": ""} ],
  "Instances": [ { "id": "", "input": "", "output": [""]} ],
}