metadata
dataset_info:
features:
- name: question
dtype: string
- name: options
struct:
- name: A
dtype: string
- name: B
dtype: string
- name: C
dtype: string
- name: D
dtype: string
- name: answer
dtype: string
- name: reasoning
dtype: string
- name: source_split
dtype: string
- name: source_idx
dtype: int64
- name: messages
list:
- name: role
dtype: string
- name: content
dtype: string
splits:
- name: train
num_bytes: 32356038
num_examples: 20665
download_size: 26572278
dataset_size: 32356038
configs:
- config_name: default
data_files:
- split: train
path: data/train-*