metadata
dataset_info:
features:
- name: critique_prompt
list:
- name: content
dtype: string
- name: role
dtype: string
- name: metadata
dtype: string
- name: model_responses__critique
sequence: string
- name: critique_prompt__critique__metadata
dtype: string
- name: model_responses__critique__metadata
dtype: string
splits:
- name: train
num_bytes: 1417024303
num_examples: 262840
download_size: 287342898
dataset_size: 1417024303
configs:
- config_name: default
data_files:
- split: train
path: data/train-*