metadata
dataset_info:
features:
- name: input
dtype: string
- name: instruction
dtype: string
- name: output
dtype: int64
splits:
- name: train
num_bytes: 4673818
num_examples: 15682
download_size: 497941
dataset_size: 4673818
configs:
- config_name: default
data_files:
- split: train
path: data/train-*