metadata
dataset_info:
features:
- name: id
dtype: int64
- name: text
dtype: string
- name: url
dtype: string
- name: perplexity_kenlm
dtype: int64
- name: label_identity_attack
dtype: float64
- name: label_insult
dtype: float64
- name: label_obscene
dtype: float64
- name: label_severe_toxicity
dtype: float64
- name: label_threat
dtype: float64
- name: label_toxicity
dtype: float64
splits:
- name: train
num_bytes: 58313946485
num_examples: 5110923
download_size: 30645534064
dataset_size: 58313946485
configs:
- config_name: default
data_files:
- split: train
path: data/train-*