metadata
dataset_info:
features:
- name: entries
list:
- name: content
dtype: string
- name: role
dtype: string
splits:
- name: train
num_bytes: 1406359674
num_examples: 5000000
download_size: 689687122
dataset_size: 1406359674
configs:
- config_name: default
data_files:
- split: train
path: data/train-*