metadata

dataset_info:
  features:
    - name: problem
      dtype: string
    - name: answer
      dtype: string
  splits:
    - name: train
      num_bytes: 8222880.032591735
      num_examples: 39701
  download_size: 4587138
  dataset_size: 8222880.032591735
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: mit
task_categories:
  - question-answering
language:
  - en
size_categories:
  - 10K<n<100K

Orz Math Filtered

Dataset from Open-Reasoner-Zero. We clean the original data by:

remove incomplete problem and answer
remove image-based problem
remove single-choice problem
remove url

The details for how to collect data are described in their paper.

Citation

@misc{OpenReasonerZero2025,
  title={Open-Reasoner-Zero: An Open Source Approach to Scaling Reinforcement Learning on the Base Model},
  author={Jingcheng Hu and Yinmin Zhang and Qi Han and Daxin Jiang and Xiangyu Zhang, Heung-Yeung Shum},
  year={2025},
  howpublished={\url{https://github.com/Open-Reasoner-Zero/Open-Reasoner-Zero}},
}