Datasets:
metadata
license: apache-2.0
task_categories:
- text-classification
- text-generation
language:
- en
tags:
- code
pretty_name: CodeNanoFix
size_categories:
- 10K<n<100K
CodeNanoFix consists of tuples of problem description, buggy code, and correct code, designed to evaluate human-written code with subtle differences.
Paper: https://arxiv.org/abs/2412.17429
problem_id: problme index;
pos: correct code;
neg: buggy code;
nl: natural language description;
Citation:
@article{liang2024condor,
title={Condor: A Code Discriminator Integrating General Semantics with Code Details},
author={Liang, Qingyuan and Zhang, Zhao and Liu, Chen and Sun, Zeyu and Zhang, Wenjie and Chen, Yizhou and Zhao, Zixiao and Luo, Qi and Wang, Wentao and Jiang, Yanjie and others},
journal={arXiv preprint arXiv:2412.17429},
year={2024}
}