metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: answer
dtype: string
- name: percent
dtype: int64
splits:
- name: train
num_bytes: 5550128
num_examples: 2000
download_size: 1524819
dataset_size: 5550128
configs:
- config_name: default
data_files:
- split: train
path: data/train-*