README.md

metadata

dataset_info:
  features:
    - name: id
      dtype: int64
    - name: text
      dtype: string
    - name: Instruction
      dtype: string
    - name: Reasoning
      dtype: string
    - name: Reasoning Answer
      dtype: string
    - name: Final Answer
      dtype: string
    - name: __index_level_0__
      dtype: int64
  splits:
    - name: train
      num_bytes: 3119512656
      num_examples: 730695
  download_size: 694203563
  dataset_size: 3119512656
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
license: cc-by-sa-3.0

kowikitext-qa-ref-detail-preview

Korean Wikipedia texts into QA-like with reasoning process dataset.

This release is preview released.

License follows original Wikipedia(CC-BY-SA-3.0)