metadata
dataset_info:
features:
- name: id
dtype: string
- name: text_type
dtype: string
- name: original_text
dtype: string
- name: cot_response
dtype: string
- name: prompt
list:
- name: content
dtype: string
- name: role
dtype: string
- name: ground_truth
dtype: string
- name: prompt_len
dtype: int64
splits:
- name: train
num_bytes: 35650785
num_examples: 10011
download_size: 16225363
dataset_size: 35650785
configs:
- config_name: default
data_files:
- split: train
path: data/train-*