metadata
dataset_info:
features:
- name: conversations
list:
- name: from
dtype: string
- name: value
dtype: string
- name: problem
dtype: string
- name: deepseek_reasoning
dtype: string
- name: deepseek_solution
dtype: string
- name: ground_truth_solution
dtype: string
- name: domain
dtype: string
- name: source
dtype: string
- name: test_cases
dtype: string
- name: starter_code
dtype: string
- name: response
dtype: string
splits:
- name: train
num_bytes: 11360748964
num_examples: 193396
download_size: 5276569583
dataset_size: 11360748964
configs:
- config_name: default
data_files:
- split: train
path: data/train-*