metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: input
dtype: string
- name: token
sequence: string
- name: answer
dtype: string
splits:
- name: train
num_bytes: 2050285.17618469
num_examples: 7900
- name: test
num_bytes: 512830.82381530985
num_examples: 1976
download_size: 539253
dataset_size: 2563116
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*