metadata
dataset_info:
features:
- name: input_ids
sequence: int32
- name: attention_mask
sequence: int8
- name: labels
sequence: int64
- name: position_ids
sequence: int64
- name: length
dtype: int64
splits:
- name: test
num_bytes: 106336749
num_examples: 4590
download_size: 23020442
dataset_size: 106336749
configs:
- config_name: default
data_files:
- split: test
path: data/test-*