metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: chosen
dtype: string
- name: rejected
dtype: string
- name: chosen_score
dtype: float64
- name: rejected_score
dtype: float64
splits:
- name: train
num_bytes: 161700273
num_examples: 92858
- name: test
num_bytes: 146776240
num_examples: 83802
download_size: 42117145
dataset_size: 308476513
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: test
path: data/test-*