Synthetic vs. Curated Security Data in Code Llama Vulnerability Detection

This report synthesises findings from 14 peer-reviewed papers addressing the following research question: What is the difference in cross-dataset generalization performance between Code Llama models augmented with synthetic code vulnerabilities versus those trained on curated security datasets when. 8 claims were extracted from source literature; 7 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 7.9/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research goal: What is the difference in cross-dataset generalization performance between Code Llama models augmented with synthetic code vulnerabilities versus those trained on curated security datasets when evaluated on diverse vulnerability detection benchmarks?

Autonomous literature synthesis. Automated review score: 7.9/10. Full text and citation available at Assignee Research.