metadata
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
dataset_info:
features:
- name: en_toxic_comment
dtype: string
- name: en_neutral_comment
dtype: string
- name: edit_ops
list:
- name: content
dtype: string
- name: operation
dtype: string
- name: position
dtype: int64
- name: replacement_content
dtype: string
splits:
- name: train
num_bytes: 4067122
num_examples: 19744
download_size: 1959427
dataset_size: 4067122