Dataset · sean0042
Korean medical multiple-choice exam benchmark
sean0042/KorMedMCQA
7,469 test questions from South Korean medical licensing exams.
What's inside
The first Korean medical multiple-choice question-answering (MCQA) benchmark based on professional licensing exams in South Korea from 2012–2024. It contains 7,469 text and table questions in Korean from exams for doctors, nurses, pharmacists, and dentists across a wide range of medical disciplines.
What you can do with it
- Evaluating medical knowledge and reasoning of language models in Korean
- Supervised fine-tuning (SFT) of specialized medical assistants
- Tasks
- Question answering
- Languages
- Korean
- Size
- 1K–10K rows
- Downloads this month
- 1.9K
- Likes
- 44
Models trained on it
- mradermacher/BioLlama-Ko-8B-GGUF 586
- tensorblock/BioLlama-Ko-8B-GGUF 247
- HyeonSang/Midm-2.0-KorMedMCQAText generation 51
- HyeonSang/Midm-2.0-KorMedMCQA-QAT-bf16Text generation 51
- csj9630/qwen3-4b-medical-qloraText generation 18
- iRASC/BioLlama-Ko-8BText generation 17
- youknowho/kanana-nano-2.1b-fft-KorMedMCQA-finalText generation 16
- ChuGyouk/ko-med-gemma-2-9b-it-merge2 15
How to load it
from datasets import load_dataset
ds = load_dataset("sean0042/KorMedMCQA")