metadata
dataset_info:
features:
- name: text
dtype: string
- name: activations
sequence: float64
splits:
- name: train
num_bytes: 41656850477
num_examples: 1141728
download_size: 2924704503
dataset_size: 41656850477
configs:
- config_name: default
data_files:
- split: train
path: data/train-*