metadata
dataset_info:
config_name: arxiv
features:
- name: content
dtype: string
splits:
- name: forget
num_bytes: 22127152
num_examples: 500
- name: approximate
num_bytes: 371246809
num_examples: 6155
- name: retain
num_bytes: 84373706
num_examples: 2000
download_size: 216767189
dataset_size: 477747667
configs:
- config_name: arxiv
data_files:
- split: forget
path: arxiv/forget-*
- split: approximate
path: arxiv/approximate-*
- split: retain
path: arxiv/retain-*