datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
MedQA-USMLE-4-options-clean
MedQA-USMLE-4-options-clean Dataset
Overview
MedQA-USMLE-4-options-clean is an enhanced medical question-answering benchmark that builds upon the MedQA-USMLE dataset. Physicians analyzed the 1373 questions in the original dataset and moved 52 questions that were either malformed or incomplete to another split incomplete.
Key Features
Relabeled malformed/incorrect questions
Dataset Details
Size: 1373
Language: English
Data Source… See the full description on the dataset page: https://huggingface.co/datasets/maximegmd/MedQA-USMLE-4-options-clean.mmlu-5-options-rl-ready
MMLU – 5-Options RL-Ready
A standardized, RL-friendly remix of MMLU with explicit negatives and a unified five-option presentation string for each question. Ideal for DPO and other RL setups while remaining drop-in for classic multiple-choice evaluation.
What’s inside
Splits & size: ~97.8k train + 2k test ≈ 99.8k total.
Schema (core fields):
question: str
choices: list[str] (canonical options, typically 4 as in original MMLU)
answer: int (0-based index)
task: str… See the full description on the dataset page: https://huggingface.co/datasets/openmed-community/mmlu-5-options-rl-ready.MedQA-USMLE-4-options-hfMedQA-USMLE-4-options_preprocess
MedQA-USMLE Preprocessed Dataset
This dataset is a preprocessed version of GBaker/MedQA-USMLE-4-options.
The data has been formatted into a question and answer structure suitable for training or evaluating instruction-following language models.
Data Structure
question: The original medical question combined with the four multiple-choice options.
answer: The correct answer index, prefixed with ####.
Example
Question:
A 60-year-old woman comes to the emergency… See the full description on the dataset page: https://huggingface.co/datasets/LLMcompe-Team-Watanabe/MedQA-USMLE-4-options_preprocess.fr-medqa-5_options