metadata
dataset_info:
features:
- name: prompt
list:
- name: content
dtype: string
- name: role
dtype: string
- name: ground_truth
dtype: string
splits:
- name: train
num_bytes: 83518067
num_examples: 180709
download_size: 32990617
dataset_size: 83518067
configs:
- config_name: default
data_files:
- split: train
path: data/train-*