README.md

metadata

dataset_info:
  features:
    - name: english
      dtype: string
    - name: korean
      dtype: string
    - name: from
      dtype: string
    - name: category
      dtype: string
    - name: en_token_len
      dtype: int64
    - name: ko_token_len
      dtype: int64
  splits:
    - name: train
      num_bytes: 464783988.3856251
      num_examples: 98712
  download_size: 243861792
  dataset_size: 464783988.3856251
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*

Parallel

• ENG: prometheus-eval/Feedback-Collection
• KOR: nayohan/feedback-collection-ko-full

Filter out 1240 sentences with repeated sentences among the translated datasets

• 99,952 -> 98,712