metadata
dataset_info:
features:
- name: prompt
dtype: string
- name: chosen
dtype: string
- name: rejected
dtype: string
- name: problem_id
dtype: string
- name: chosen_explanation
dtype: string
- name: rejected_explanation
dtype: string
- name: canonical_solution
dtype: string
splits:
- name: train
num_bytes: 30225
num_examples: 25
download_size: 25241
dataset_size: 30225
configs:
- config_name: default
data_files:
- split: train
path: data/train-*