What type of study is this?

This is a Quantitative Study study.

September 23, 2025

Smart Contracts- Vulnerabilities, CodeLlama Usage and Gas Driven Detection

NKNatan KatzUniversidade Federal do Rio Grande do Sul NKNatan KatzUniversidade Federal do Rio Grande do Sul

Key Points

The study finds two effective methods for detecting malicious contracts, utilizing both opcode and Solidity source to analyze vulnerabilities.
An XGBOOST model is developed, showing how unusual use of contract functions and gas properties can identify malicious transactions.
Using a finetuned CodeLlama allows for improved understanding of smart contract vulnerabilities, enhancing detection capabilities.
Detection models were built based on assumptions linking function usage with malicious behavior, providing insights for future security measures.

Abstract

Smart contracts are a major tool in Ethereum transactions. Therefore, hackers can exploit them by generating malicious code and using their vulnerabilities to perform malicious transactions. This paper presents two successful approaches for detecting malicious contracts: one uses opcode and relies on GPT2, and the other uses the Solidity source and a LORA finetuned CodeLlama. Finally, we present an XGBOOST model that combines gas properties and Hexadecimal signatures for detecting malicious transactions. This approach relies on early assumptions that maliciousness is manifested by the uncommon usage of the contracts’ functions and the effort to pursue the transaction.

AIに質問

Bookmark

View Full Paper