metadata
dataset_info:
features:
- name: INSTRUCTION
dtype: string
- name: RESPONSE
dtype: string
- name: SOURCE
dtype: string
splits:
- name: train
num_bytes: 7600144
num_examples: 4747
download_size: 3336556
dataset_size: 7600144
configs:
- config_name: default
data_files:
- split: train
path: data/train-*