metadata
dataset_info:
features:
- name: text
dtype: string
- name: toxic
dtype: float32
- name: severe_toxic
dtype: float32
- name: obscene
dtype: float32
- name: threat
dtype: float32
- name: insult
dtype: float32
- name: identity_attack
dtype: float32
splits:
- name: train
num_bytes: 218203977
num_examples: 1530625
download_size: 185014413
dataset_size: 218203977
configs:
- config_name: default
data_files:
- split: train
path: data/train-*