SecVulEval provides a statement-level C/C++ vulnerability benchmark with context; state-of-the-art LLMs achieve only 23.83% F1 on locating vulnerable statements with correct reasoning.
Diversevul: A new vulnerable source code dataset for deep learning based vulnerability detection
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
dataset 1
citation-polarity summary
fields
cs.SE 1years
2025 1verdicts
CONDITIONAL 1roles
dataset 1polarities
contest 1representative citing papers
citing papers explorer
-
SecVulEval: Benchmarking LLMs for Real-World C/C++ Vulnerability Detection
SecVulEval provides a statement-level C/C++ vulnerability benchmark with context; state-of-the-art LLMs achieve only 23.83% F1 on locating vulnerable statements with correct reasoning.