metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: output
sequence: string
- name: scores
sequence: float64
splits:
- name: train
num_bytes: 21664659251
num_examples: 61135
download_size: 7570708794
dataset_size: 21664659251
configs:
- config_name: default
data_files:
- split: train
path: data/train-*