Datasets:
metadata
license: apache-2.0
task_categories:
- question-answering
tags:
- medical
- usmle
- multiple-choice
USMLE Questions Dataset
Evaluation dataset of USMLE-style multiple choice questions.
Dataset structure
Each row has:
- question: Question statement
- options: List of options (e.g.
["(A) ...", "(B) ...", ...]) - correct_answer: Correct option letter (A–E)
Loading
from datasets import load_dataset
ds = load_dataset("prapaa/usmle", data_files="{"train": "train.jsonl"}, split="train")