metadata
dataset_info:
features:
- name: index
dtype: int64
- name: conversations
list:
- name: content
dtype: string
- name: role
dtype: string
splits:
- name: train
num_bytes: 23500885
num_examples: 2000
download_size: 5655265
dataset_size: 23500885
configs:
- config_name: default
data_files:
- split: train
path: data/train-*