metadata
dataset_info:
features:
- name: problem
dtype: string
- name: reasoning
dtype: string
- name: deepseek_solution
dtype: string
- name: ground_truth_solution
dtype: string
- name: correct
dtype: bool
- name: judge_reasoning
dtype: string
splits:
- name: train
num_bytes: 2374426933.329145
num_examples: 89120
download_size: 893744769
dataset_size: 2374426933.329145
configs:
- config_name: default
data_files:
- split: train
path: data/train-*