Fine-tuned decoder-only LLMs fall into a Semantic Trap on vulnerability detection, achieving high scores on unpaired normal code but failing on paired vulnerable-patched code, semantic perturbations, and gap analysis, while reasoning supervision reduces symptoms at the cost of recall.
Title resolution pending
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
fields
cs.CR 2verdicts
UNVERDICTED 2representative citing papers
PMDetector is a hybrid static-plus-LLM framework that detects price manipulation in DeFi protocols via taint analysis, defense filtering, attack simulation, and validation, achieving 88% precision and 90% recall on 73 vulnerable plus 288 benign contracts.
citing papers explorer
-
Do Fine-Tuned LLMs Understand Vulnerabilities? An Investigation into the Semantic Trap
Fine-tuned decoder-only LLMs fall into a Semantic Trap on vulnerability detection, achieving high scores on unpaired normal code but failing on paired vulnerable-patched code, semantic perturbations, and gap analysis, while reasoning supervision reduces symptoms at the cost of recall.
-
LLM-Powered Detection of Price Manipulation in DeFi
PMDetector is a hybrid static-plus-LLM framework that detects price manipulation in DeFi protocols via taint analysis, defense filtering, attack simulation, and validation, achieving 88% precision and 90% recall on 73 vulnerable plus 288 benign contracts.