BriefGPT.xyz
Mar, 2025
评估LLaMA 3.2在软件漏洞检测中的表现
Evaluating LLaMA 3.2 for Software Vulnerability Detection
HTML
PDF
José Gonçalves, Miguel Silva, Bernardo Cabral, Tiago Dias, Eva Maia...
TL;DR
本研究解决了深度学习在软件漏洞检测中对大量真实数据需求的问题,提出了经过优化的DiverseVul数据集,以提供高质量的训练样本。通过对LLaMA 3.2模型的微调,实验结果显示处理技术显著提高了性能,F1得分达66%,显著优于基线的47%。
Abstract
Deep Learning
(DL) has emerged as a powerful tool for
Vulnerability Detection
, often outperforming traditional solutions. However, developing effective DL models requires large amounts of real-world data, which c
→