This research introduces a novel approach to enhancing smart contract security by leveraging fine-tuned Large Language Models to detect vulnerabilities traditional methods miss.

• Created a comprehensive dataset of 215 real-world DApp projects (4,998 contracts)
• Developed a technique to identify complex vulnerabilities including hard-to-detect logical errors
• Demonstrated superior performance in detecting emerging and machine-unauditable flaws
• Provides a practical solution for blockchain security enhancement

This advancement is crucial for the blockchain ecosystem as smart contract vulnerabilities have led to billions in losses. By improving detection capabilities, this approach helps protect digital assets and builds trust in decentralized applications.

Enhancing Smart Contract Vulnerability Detection in DApps Leveraging Fine-Tuned LLM