ReVD uses synthetic vulnerability reasoning data and curriculum preference optimization to boost LLM vulnerability detection accuracy by 12-22% over prior baselines on PrimeVul and SVEN.
Specifically, line 16 of the target code does not check if the input tensors are empty before proceeding with the division
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.AI 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Boosting Vulnerability Detection of LLMs via Curriculum Preference Optimization with Synthetic Reasoning Data
ReVD uses synthetic vulnerability reasoning data and curriculum preference optimization to boost LLM vulnerability detection accuracy by 12-22% over prior baselines on PrimeVul and SVEN.