metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: context
dtype: string
- name: response
dtype: string
splits:
- name: train
num_bytes: 2131234
num_examples: 2450
download_size: 179543
dataset_size: 2131234
configs:
- config_name: default
data_files:
- split: train
path: data/train-*