Source Code Vulnerability Analysis Using GPT-2
Published in Advances in Information Security, Privacy, and Ethics • Jul 17, 2024
Authors:
Attila Magyar
Abstract
This study introduces a deep learning methodology aimed at identifying security vulnerabilities present in C/C++ source code files. Employing default tokenizers and encoders of the GPT-2 architecture, the approach involves pre-training and subsequent fine-tuning of a GPT-2 large language model using...
Finding related papers...
Discussions
(0)No comments yet
Be the first to share your thoughts!