metadata
dataset_info:
features:
- name: inputs_embeds
sequence: float64
- name: text
dtype: string
- name: reward
dtype: float64
splits:
- name: train
num_bytes: 708167877
num_examples: 21446
- name: test
num_bytes: 177058474
num_examples: 5362
download_size: 166305886
dataset_size: 885226351
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*