PrimeVul
A benchmark dataset for vulnerability detection research containing vulnerability samples and annotations from real-world code, commonly used to evaluate the accuracy and generalization ability of vulnerability detection models.
Timeline (last 90 days)
On the PrimeVul held-out test set, REVA with GLM-5.3-Flash achieves 60.0% accuracy at about one second per function.
On PrimeVul, the agent-style detector VulTriage (also running on GLM-5.3-Flash) achieves 59.2% accuracy.
PrimeVul is a function-level vulnerability detection benchmark that uses stricter temporal splitting and deduplication/leak-removal processing to avoid duplicate or highly similar samples between the training and test sets.
REVA's results are concentrated on only one benchmark, PrimeVul; the accuracy of around 60% is still far from perfect, and probe quality depends on the coverage of the training data, while threat models and static analysis quality also propagate to the final judgment.
All Facts (4)
On the PrimeVul held-out test set, REVA with GLM-5.3-Flash achieves 60.0% accuracy at about one second per function.
50%UnverifiedOn PrimeVul, the agent-style detector VulTriage (also running on GLM-5.3-Flash) achieves 59.2% accuracy.
50%UnverifiedPrimeVul is a function-level vulnerability detection benchmark that uses stricter temporal splitting and deduplication/leak-removal processing to avoid duplicate or highly similar samples between the training and test sets.
50%UnverifiedREVA's results are concentrated on only one benchmark, PrimeVul; the accuracy of around 60% is still far from perfect, and probe quality depends on the coverage of the training data, while threat models and static analysis quality also propagate to the final judgment.
50%