智能合约漏洞检测数据集
收藏资源简介:
该数据集由海南大学网络空间安全学院创建,包含215个真实世界的DApp项目,共4998个智能合约。数据集旨在用于智能合约漏洞检测,特别是针对机器难以审核的逻辑错误,如代币价格操纵等。数据集的构建通过从Code4rena和Slowmist平台收集审计过的合同项目,并使用SmartCollect工具获取完整的DApp项目依赖合同。数据集还采用了随机过采样技术以解决数据不平衡问题,以提高模型的检测性能。
This dataset was developed by the School of Cyberspace Security, Hainan University. It contains 215 real-world DApp projects, with a total of 4998 smart contracts. The dataset is designed for smart contract vulnerability detection, especially targeting logic errors that are difficult for machines to audit, such as token price manipulation. The dataset was constructed by collecting audited contract projects from the Code4rena and Slowmist platforms, and utilizing the SmartCollect tool to acquire complete dependent contracts of DApp projects. Additionally, random oversampling technology was adopted to resolve the data imbalance problem, thereby enhancing the detection performance of models.

- 1Enhancing Smart Contract Vulnerability Detection in DApps Leveraging Fine-Tuned LLM海南大学网络空间安全学院 · 2025年



