Dataset · qiaojin
Medical Yes/No Question Answering Dataset
qiaojin/PubMedQA
Medical research questions with yes/no/maybe answers based on PubMed abstracts.
What's inside
PubMedQA offers answering medical research questions with yes/no/maybe options using corresponding abstracts. In the labeled pqa_labeled subset, 500 questions serve as a test set. The data is in English in text format, with an official leaderboard.
What you can do with it
- Training a QA model on biomedical texts
- Testing LLMs on yes/no/maybe answers
- Evaluating retrieval and QA on medical abstracts
- Tasks
- Question answering
- Languages
- English
- Size
- 100K–1M rows
- Downloads this month
- 41.9K
- Likes
- 360
Models trained on it
- StrandsAgents/strands-decider-2B-qwen3.5-v1-2610Text classification 25K
- StrandsAgents/strands-decider-2B-hobson-v21Text classification 15.3K
- StrandsAgents/strands-decider-2B-hobson-v19Text classification 7K
- mradermacher/eubiota-planner-8b-i1-GGUFReinforcement learning 2.2K
- mradermacher/Sanad-1.0-i1-GGUF 2K
- mradermacher/Medra-GGUFQuestion answering 1.9K
- mradermacher/Mistral-Nemo-NT-Ko-12B-sft-i1-GGUF 1.8K
- mradermacher/MedLLaMA-3-GGUF 1.7K
How to load it
from datasets import load_dataset
ds = load_dataset("qiaojin/PubMedQA")