Le Qin

A Review of Source Code Vulnerability Detection Methods Based on Large Language Models: Models, Techniques, and Evaluation

In source code vulnerability detection, large language models (LLMs) are no longer used only as substitutes for traditional fixed classifiers. With prompt design, parameter fine-tuning, and program-context information, they can judge whether code contains vulnerabilities and can also provide suspicious locations and ex...