Instruction-Finetuned Small Language Models vs. Domain-Adapted LLaVul in Zero-Shot CWE Detection

This report synthesises findings from 8 peer-reviewed papers addressing the following research question: How does the zero-shot transfer performance of fine-tuned SLMs compare to domain-adapted LLaVul models when evaluating CWE detection in non-Python languages (e.g., Java, C) on the TUFano or. Finetuning language models on a collection of datasets phrased as instructions has been shown to improve model performance and generalization to unseen tasks. In this paper we explore instruction finetuning with a particular focus on (1) scaling the number of tasks, (2) scaling. 6 claims were extracted from source literature; 6 were independently verified against retrieved documents. An automated multi-reviewer quality assessment produced a score of 9.2/10. This report is a machine-generated literature synthesis and does not constitute original research.

Research goal: How does the zero-shot transfer performance of fine-tuned SLMs compare to domain-adapted LLaVul models when evaluating CWE detection in non-Python languages (e.g., Java, C) on the TUFano or VulDeePecker benchmarks?

Autonomous literature synthesis. Automated review score: 9.2/10. Full text and citation available at Assignee Research.