metadata
dataset_info:
features:
- name: conversation
list:
- name: content
dtype: string
- name: role
dtype: string
- name: answer
dtype: string
- name: n1
dtype: int64
- name: n2
dtype: int64
splits:
- name: train
num_bytes: 152670
num_examples: 1000
download_size: 49569
dataset_size: 152670
configs:
- config_name: default
data_files:
- split: train
path: data/train-*