Interpretable Vulnerability Detection in LLMs: A BERT-Based Approach with SHAP Explanations
Source abstract
Abstract
Source code vulnerabilities present significant security threats, necessitating effective detection techniques. Rigid rule-sets and pattern matching are the foundation of traditional static analysis tools, which drown develop... | Find, read and cite all the research you need on Tech Science Press
Source: openalex_inverted_index · Quality: baseline
Research context
OpenAlex topics: Web Application Security Vulnerabilities · Information and Cyber Security
SciVX discovery tags are terminology-based. OpenAlex topics are source-provided classifications.
References (23)
Show source-provided references
1. Generative AI in cybersecurity: a comprehensive review of LLM applications and vulnerabilities10.1016/j.iotcps.2025.01.001 ↗
2. Large language model for vulnerability detection and repair: literature review and the road ahead
3. Context unlocks emotions: text-based emotion classification dataset auditing with large language models
4. Instance-level weighted contrast learning for text classification10.3390/app15084236 ↗
5. Using an LLM to help with code understanding
6. Explainability for large language models: a survey
7. Devlin J, Chang M-W, Lee K, Toutanova K. BERT: pre-training of deep bidirectional transformers for language understanding. arXiv:1810.04805. 2019.
8. When large language models meet personalization: perspectives of challenges and opportunities10.1007/s11280-024-01276-1 ↗
9. Mathews NS, Brus Y, Aafer Y, Nagappan M, McIntosh S. LLbezpeky: leveraging large language models for vulnerability detection. arXiv:2401.01269. 2024. doi:10.48550/arxiv.2401.01269.
10. Gao Z, Wang H, Zhou Y, Zhu W, Zhang C. How far have we gone in vulnerability detection using large language models. arXiv:2311.12420. 2023. doi:10.48550/arxiv.2311.12420.
11. Software vulnerability detection using large language models
12. CodeT5: identifier-aware unified pre-trained encoder-decoder models for code understanding and generation
13. Large language model for vulnerability detection: emerging results and future directions
14. Sun Y, Wu D, Xue Y, Liu H, Ma W, Zhang L et al. LLM4Vuln: a unified evaluation framework for decoupling and enhancing LLMs’ vulnerability reasoning. arXiv.2401.16185. 2024. doi:10.48550/arxiv.2401.16185. .
15. LLMs cannot reliably identify and reason about security vulnerabilities (Yet?): a comprehensive evaluation, framework, and benchmarks
16. Towards an understanding of large language models in software engineering tasks10.1007/s10664-024-10602-0 ↗
17. Shui X, Ru Z. Bridging the gap between explainability and large language models; 2025 [cited 2025 Jun 6]. Available from: https://hal.science/hal-05011844.
18. How explainable AI affects human performance: a systematic review of the behavioural consequences of saliency maps10.1080/10447318.2024.2381929 ↗
19. The proposal of undersampling method for learning from imbalanced datasets10.1016/j.procs.2019.09.167 ↗
20. Machine learning with oversampling and undersampling techniques: overview study and experimental results
21. Adaptive biased stochastic optimization10.1109/TPAMI.2025.3528193 ↗
22. Basheer · Artificial intelligence applications and innovations · 2024 · 316
23. DiverseVul: a new vulnerable source code dataset for deep learning based vulnerability detection