AdaptLLM/med_knowledge_prob

Biomedicine Knowledge Probing数据集是一个用于医学领域知识探测的数据集，包含多个医学相关领域的测试数据，如麻醉学、解剖学、生物化学等。该数据集用于文本分类、问答和零样本分类任务，旨在通过持续预训练在特定领域语料库上适应大型语言模型，提升其在生物医学、金融和法律领域的提示性能。The Biomedicine Knowledge Probing dataset is a dataset designed for knowledge probing in the medical domain, which contains test data across multiple medicine-related fields including anesthesiology, anatomy, biochemistry, and others. This dataset supports tasks such as text classification, question answering, and zero-shot classification. It aims to adapt large language models (LLMs) via continued pre-training on domain-specific corpora, and improve their prompting performance in the biomedicine, finance, and legal fields.

原始信息汇总

数据集概述

数据集配置

• config_name: 数据集的名称，涵盖多个医学领域。
• data_files: 每个配置下的数据文件列表。
  • split: 数据集的分割类型，此处均为测试集。
  • path: 数据文件的路径，格式为.jsonl。

任务类别

• text-classification
• question-answering
• zero-shot-classification

语言

• en

标签

• medical
• chemistry
• biology

数据文件列表

• Anaesthesia: Anaesthesia.jsonl
• Anatomy: Anatomy.jsonl
• Biochemistry: Biochemistry.jsonl
• Dental: Dental.jsonl
• ENT: ENT.jsonl
• Forensic Medicine: Forensic Medicine.jsonl
• Gynaecology & Obstetrics: Gynaecology & Obstetrics.jsonl
• Medicine: Medicine.jsonl
• Microbiology: Microbiology.jsonl
• Ophthalmology: Ophthalmology.jsonl
• Orthopedics: Orthopedics.jsonl
• Pathology: Pathology.jsonl
• Pediatrics: Pediatrics.jsonl
• Pharmacology: Pharmacology.jsonl
• Physiology: Physiology.jsonl
• Psychiatry: Psychiatry.jsonl
• Radiology: Radiology.jsonl
• Skin: Skin.jsonl
• Social & Preventive Medicine: Social & Preventive Medicine.jsonl
• Surgery: Surgery.jsonl
• Unknown: Unknown.jsonl