metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: token_length
dtype: int64
- name: system_prompt
dtype: string
splits:
- name: train
num_bytes: 84337
num_examples: 20
download_size: 74831
dataset_size: 84337
configs:
- config_name: default
data_files:
- split: train
path: data/train-*