metadata
dataset_info:
features:
- name: token_positions
sequence:
sequence: int64
- name: token_id
dtype: int64
splits:
- name: validation
num_bytes: 112719948
num_examples: 3719
download_size: 20085684
dataset_size: 112719948
configs:
- config_name: default
data_files:
- split: validation
path: data/validation-*