llamafactory/PubMedQA

This dataset is a subset of https://github.com/pubmedqa/pubmedqa.

The training set is composed of 10,000 examples of the PQA-U dataset and the test set is composed of 1,000 examples of the PQA-L dataset.

Citation Information

@inproceedings{jin2019pubmedqa,
  title={PubMedQA: A Dataset for Biomedical Research Question Answering},
  author={Jin, Qiao and Dhingra, Bhuwan and Liu, Zhengping and Cohen, William and Lu, Xinghua},
  booktitle={Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP)},
  pages={2567--2577},
  year={2019}
}