metadata
dataset_info:
features:
- name: prompt
list:
- name: content
dtype: string
- name: role
dtype: string
- name: reference_answer
dtype: string
- name: data_source
dtype: string
splits:
- name: train
num_bytes: 79355639
num_examples: 184544
download_size: 30356684
dataset_size: 79355639
configs:
- config_name: default
data_files:
- split: train
path: data/train-*