Summary:
- This paper introduces "LLM-based Automated Vulnerability Detection," a framework designed to identify security flaws in source code using Large Language Models.
- The research presents a novel dataset and evaluation methodology to improve the precision of automated static analysis tools, significantly reducing false positives in software security auditing.