Datasets:

leukas
/

hatexplain-filtered

dataset_info:
  features:
    - name: text
      dtype: string
    - name: label
      dtype: bool
  splits:
    - name: train
      num_bytes: 541553
      num_examples: 4000
    - name: validation
      num_bytes: 261548
      num_examples: 1922
    - name: test
      num_bytes: 134794
      num_examples: 1000
  download_size: 634837
  dataset_size: 937895
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
      - split: validation
        path: data/validation-*
      - split: test
        path: data/test-*

Hatexplain-filtered

This is a sample of the hatexplain dataset, with most of the bad words replaced with gibberish.