metadata
dataset_info:
features:
- name: sequence_id
dtype: int64
- name: nl_scores
dtype: float32
splits:
- name: standard
num_bytes: 1757184000
num_examples: 146432000
- name: deduped
num_bytes: 1757184000
num_examples: 146432000
download_size: 2547384724
dataset_size: 3514368000
configs:
- config_name: default
data_files:
- split: standard
path: data/standard-*
- split: deduped
path: data/deduped-*