metadata
dataset_info:
features:
- name: resid_streams
sequence:
sequence: float32
splits:
- name: train
num_bytes: 442948000
num_examples: 1000
download_size: 443091322
dataset_size: 442948000
configs:
- config_name: default
data_files:
- split: train
path: data/train-*