metadata
dataset_info:
features:
- name: messages
list:
- name: role
dtype: string
- name: content
dtype: string
- name: topic
dtype: string
- name: difficulty
dtype: string
splits:
- name: train
num_bytes: 11647450
num_examples: 2000
download_size: 11603227
dataset_size: 11647450
configs:
- config_name: default
data_files:
- split: train
path: data/train-*