metadata
dataset_info:
features:
- name: problem_id
dtype: string
- name: name
dtype: string
- name: problem
dtype: string
- name: test_cases
dtype: string
- name: difficulty
dtype: string
- name: language
sequence: string
- name: source
dtype: string
- name: num_solutions
dtype: int64
- name: starter_code
dtype: string
- name: subset
dtype: string
- name: reasoning
dtype: string
- name: deepseek_solution
dtype: string
- name: correctness
dtype: bool
- name: reason
dtype: string
- name: reasoning_tokens
dtype: int64
- name: solution_tokens
dtype: int64
- name: total_tokens
dtype: int64
- name: problem_tokens
dtype: int64
- name: conversations
list:
- name: content
dtype: string
- name: role
dtype: string
- name: system
dtype: string
- name: formatted_text
dtype: string
- name: total_templated_tokens
dtype: int64
splits:
- name: train
num_bytes: 6885316933
num_examples: 28338
download_size: 3226091210
dataset_size: 6885316933
configs:
- config_name: default
data_files:
- split: train
path: data/train-*