metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: chosen
dtype: string
- name: chosen_score
dtype: float64
- name: rejected
dtype: string
- name: rejected_score
dtype: float64
splits:
- name: train
num_bytes: 26507779
num_examples: 3410
- name: validation
num_bytes: 1398384
num_examples: 183
download_size: 18743233
dataset_size: 27906163
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
- split: validation
path: data/validation-*