metadata
dataset_info:
features:
- name: problem_spec
dtype: string
- name: correct_answer
dtype: string
- name: corrupted_answer
dtype: string
- name: tests
dtype: string
- name: applied_mutations
list: string
splits:
- name: train
num_bytes: 547340877.4246002
num_examples: 5442
download_size: 336972150
dataset_size: 547340877.4246002
configs:
- config_name: default
data_files:
- split: train
path: data/train-*