REVIEW 9 cited by
Understanding the Effectiveness of Large Language Models in Detecting Security Vulnerabilities
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
read the original abstract
While automated vulnerability detection techniques have made promising progress in detecting security vulnerabilities, their scalability and applicability remain challenging. The remarkable performance of Large Language Models (LLMs), such as GPT-4 and CodeLlama, on code-related tasks has prompted recent works to explore if LLMs can be used to detect vulnerabilities. In this paper, we perform a more comprehensive study by concurrently examining a higher number of datasets, languages and LLMs, and qualitatively evaluating performance across prompts and vulnerability classes while addressing the shortcomings of existing tools. Concretely, we evaluate the effectiveness of 16 pre-trained LLMs on 5,000 code samples from five diverse security datasets. These balanced datasets encompass both synthetic and real-world projects in Java and C/C++ and cover 25 distinct vulnerability classes. Overall, LLMs across all scales and families show modest effectiveness in detecting vulnerabilities, obtaining an average accuracy of 62.8% and F1 score of 0.71 across datasets. They are significantly better at detecting vulnerabilities only requiring intra-procedural analysis, such as OS Command Injection and NULL Pointer Dereference. Moreover, they report higher accuracies on these vulnerabilities than popular static analysis tools, such as CodeQL. We find that advanced prompting strategies that involve step-by-step analysis significantly improve performance of LLMs on real-world datasets in terms of F1 score (by upto 0.18 on average). Interestingly, we observe that LLMs show promising abilities at performing parts of the analysis correctly, such as identifying vulnerability-related specifications and leveraging natural language information to understand code behavior (e.g., to check if code is sanitized). We expect our insights to guide future work on LLM-augmented vulnerability detection systems.
Forward citations
Cited by 9 Pith papers
-
A Mixture of Linear Corrections Generates Secure Code
An inference-time mixture of linear correction vectors, derived from linear probes on LLM hidden states, improves the security and functionality of code generated by Qwen2.5-Coder and CodeLlama models.
-
Measuring and Augmenting Large Language Models for Solving Capture-the-Flag Challenges
A benchmark and agent for CTF solving, but the agent's retrieval database appears to contain the answers to the test challenges, undermining the reported improvements.
-
LLM Embedding-based Attribution (LEA): Quantifying Source Contributions to Generative Model's Response for Vulnerability Analysis
LEA uses rank-based linear dependence of layer-0 hidden states to attribute each response token to query, retrieved context, or internal knowledge, and distinguishes valid from generic retrieval with over 95% accuracy.
-
Mono: Is Your "Clean" Vulnerability Dataset Really Solvable? Exposing and Trapping Undecidable Patches and Beyond
Mono reports that 31% of MegaVul patches are non-security and about 16.7% of CVEs are 'undecidable', while its added context raises LLM vulnerability detection F1 by up to 15%.
-
DCE-LLM: Dead Code Elimination with Large Language Models
DCE-LLM reports over 94% F1 for unused and unreachable code detection using a CodeBERT pivot with leave-one-out attribution and a Qwen2-7B fine-tuned for explanations and patches.
-
SV-TrustEval-C: Evaluating Structure and Semantic Reasoning in Large Language Models for Source Code Vulnerability Analysis
Large language models perform poorly on a new C-code vulnerability benchmark, indicating they rely on pattern matching rather than genuine reasoning.
-
Adaptive Plan-Execute Framework for Smart Contract Security Auditing
SmartAuditFlow claims 100% detection on a standard smart contract benchmark and all 13 tested CVEs via a plan-execute LLM workflow, though the supporting evaluation has major reproducibility and validation gaps.
-
Large Language Models in Cybersecurity: Applications, Vulnerabilities, and Defense Techniques
A survey that maps LLM applications, vulnerabilities, and defenses across eight cybersecurity domains, but with significant citation and rigor problems.
-
LLMs in Software Security: A Survey of Vulnerability Detection Techniques and Insights
A survey of LLM-based vulnerability detection covering 58 papers, with a taxonomy, dataset overview, and gap analysis, but limited by non-transparent selection and unsupported quantitative claims.
Discussion (0). Continue with ORCID to comment.