Thi-Thu-Huong/Balanced-Ethereum-Smart-Contract

平衡以太坊智能合约数据集是一个用于检测以太坊智能合约中安全漏洞的数据集。它通过使用经过微调的基于变换器编码器模型的大型语言模型，对Solidity代码进行分析以识别潜在的安全缺陷。该数据集汇集了多个来源的数据，包括SB Curate、Benmark Solidity Smart Contract和ScrawlID，旨在确保模型能够在不同类型的漏洞上表现出色。数据集经过过采样处理以解决类别不平衡问题，从而提高训练结果的可靠性。

The Balanced Ethereum Smart Contract dataset is designed for detecting security vulnerabilities in Ethereum smart contracts. It utilizes fine-tuned transformer encoder models based on Large Language Models (LLMs) to analyze Solidity code for identifying potential security flaws. The dataset aggregates data from multiple sources, including SB Curate, Benmark Solidity Smart Contract, and ScrawlID, ensuring the models performance across different vulnerability types. The dataset has been processed with over-sampling techniques to address class imbalances, enhancing the reliability of training outcomes.