metadata
dataset_info:
features:
- name: problem_id
dtype: string
- name: task_type
dtype: string
- name: prompt
dtype: string
- name: verification_info
dtype: string
- name: metadata
dtype: string
- name: responses
sequence: string
- name: response_lens
sequence: int64
- name: avg_reward
dtype: float64
splits:
- name: train
num_bytes: 409921939.325
num_examples: 2000
download_size: 147807294
dataset_size: 409921939.325
configs:
- config_name: default
data_files:
- split: train
path: data/train-*