metadata
dataset_info:
features:
- name: row_id
dtype: string
- name: messages
list:
- name: role
dtype: string
- name: content
dtype: string
- name: thinking
dtype: string
- name: reasoning
dtype: string
- name: choice
dtype: string
- name: correct
dtype: 'null'
splits:
- name: train
num_bytes: 112110
num_examples: 5
download_size: 121563
dataset_size: 112110
configs:
- config_name: default
data_files:
- split: train
path: data/train-*