metadata
dataset_info:
features:
- name: meta_info
dtype: string
- name: question
dtype: string
- name: answer_idx
dtype: string
- name: answer
dtype: string
- name: options
list:
- name: key
dtype: string
- name: value
dtype: string
- name: responses
list:
list:
- name: content
dtype: string
- name: responses
sequence: string
- name: role
dtype: string
- name: rewards
sequence: int64
splits:
- name: train
num_bytes: 92307848.95608175
num_examples: 8475
download_size: 38217974
dataset_size: 92307848.95608175
configs:
- config_name: default
data_files:
- split: train
path: data/train-*