metadata
dataset_info:
features:
- name: prompt
list:
- name: content
dtype: string
- name: role
dtype: string
- name: output
dtype: int64
splits:
- name: train
num_bytes: 15250130
num_examples: 3818
download_size: 5969359
dataset_size: 15250130
configs:
- config_name: default
data_files:
- split: train
path: data/train-*