metadata
dataset_info:
features:
- name: question_id
dtype: int64
- name: prompt
dtype: string
- name: response_a
dtype: string
- name: response_b
dtype: string
- name: is_response_0_safe
dtype: bool
- name: is_response_1_safe
dtype: bool
- name: better_response_id
dtype: int64
- name: safer_response_id
dtype: int64
- name: better_response
dtype: string
- name: safer_response
dtype: string
- name: ranked_responses
sequence: string
splits:
- name: train
num_bytes: 1418163
num_examples: 1000
download_size: 736108
dataset_size: 1418163
configs:
- config_name: default
data_files:
- split: train
path: data/train-*