metadata
dataset_info:
features:
- name: instruction
dtype: string
- name: initial response
dtype: string
- name: question
dtype: string
- name: critique
dtype: string
- name: refined response
dtype: string
splits:
- name: train
num_bytes: 34387
num_examples: 10
download_size: 40914
dataset_size: 34387
configs:
- config_name: default
data_files:
- split: train
path: data/train-*