LLM-AuthorBench
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/LLMauthorbench/
下载链接
链接失效反馈官方服务：
资源简介：
该数据集是一个包含32,000个可编译C程序的基准测试集，这些程序由八种最先进的LLM（大型语言模型）在不同任务中生成。该数据集旨在评估LLM代码作者归属模型的性能，其中包括由GPT-4.1和GPT-4o等模型生成的代码。该数据集的任务是对C程序进行作者归属分析。
This dataset is a benchmark collection containing 32,000 compilable C programs generated by eight state-of-the-art Large Language Models (LLMs) across diverse tasks. It is designed to evaluate the performance of LLM-based code author attribution models, including code generated by models such as GPT-4.1 and GPT-4o. The core task of this dataset is author attribution analysis for C programs.
