metadata
language:
- en
dataset_info:
features:
- name: task_id
dtype: string
- name: prompt
dtype: string
- name: canonical_solution
dtype: string
- name: test
dtype: string
- name: declaration
dtype: string
- name: example_test
dtype: string
- name: full_code
dtype: string
splits:
- name: train
num_bytes: 335751
num_examples: 160
download_size: 120383
dataset_size: 335751
configs:
- config_name: default
data_files:
- split: train
path: data/train-*
Description
task id: [10, 32, 38, 50] removed from original dataset as they are buggy (not formatted accordingly) \n
added a new column full_code which contains full c++ code with header imports