Empirical study across five LLMs and four languages finds security-aware prompting changes CWE category distributions but yields no statistically significant reduction in vulnerability frequency or density.
CoRR abs/2312.05275, 2312.05275
2 Pith papers cite this work, alongside 2 external citations. Polarity classification is still indexing.
2
Pith papers citing it
2
external citations · external index
citation-role summary
background 1
citation-polarity summary
fields
cs.CR 2years
2026 2verdicts
UNVERDICTED 2roles
background 1polarities
background 1representative citing papers
SAGE uses sparse autoencoders to boost vulnerability signals in LLMs, raising internal SNR 12.7x and delivering up to 318% MCC gains on vulnerability detection benchmarks.
citing papers explorer
-
An Empirical Evaluation of LLM-Generated Code Security Across Prompting Methods
Empirical study across five LLMs and four languages finds security-aware prompting changes CWE category distributions but yields no statistically significant reduction in vulnerability frequency or density.
-
SAGE: Signal-Amplified Guided Embeddings for LLM-based Vulnerability Detection
SAGE uses sparse autoencoders to boost vulnerability signals in LLMs, raising internal SNR 12.7x and delivering up to 318% MCC gains on vulnerability detection benchmarks.