R+R: Security Vulnerability Dataset Quality Is Critical

Anurag Swarnim Yadav, Joseph N. Wilson

Large Language Models (LLMs) are of great interest in vulnerability detection and repair. The effectiveness of these models hinges on the quality of the datasets used for both trai…