metadata
dataset_info:
features:
- name: prompt_messages
list:
- name: content
dtype: string
- name: role
dtype: string
- name: answers
sequence: string
- name: gt
dtype: string
- name: rewards
sequence: float64
splits:
- name: train
num_bytes: 189878
num_examples: 34
download_size: 75068
dataset_size: 189878
configs:
- config_name: default
data_files:
- split: train
path: data/train-*