metadata
dataset_info:
features:
- name: resid_streams
sequence:
sequence: float32
splits:
- name: train
num_bytes: 79980000
num_examples: 1000
download_size: 80484833
dataset_size: 79980000
configs:
- config_name: default
data_files:
- split: train
path: data/train-*