metadata
dataset_info:
features:
- name: text
dtype: string
- name: representation
sequence: float64
- name: instruction
dtype: string
splits:
- name: train
num_bytes: 3093774132.0086627
num_examples: 1758317
- name: validation
num_bytes: 89013321.92012125
num_examples: 50472
download_size: 3476016218
dataset_size: 3182787453.928784
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*