metadata
dataset_info:
features:
- name: messages
list:
- name: role
dtype: string
- name: content
dtype: string
- name: source_dataset
dtype: string
- name: source_split
dtype: string
splits:
- name: train
num_bytes: 24025522
num_examples: 5922
download_size: 23758676
dataset_size: 24025522
configs:
- config_name: default
data_files:
- split: train
path: data/train-*