Aggregated Benchmark; Reentrancy Scenarios Dataset (RSD)
收藏资源简介:
本研究构建了两个互补的数据集:Aggregated Benchmark和Reentrancy Scenarios Dataset (RSD)。Aggregated Benchmark由威尼斯大学等机构创建,整合了432条来自先前学术数据集的真实智能合约,经过多评审员检查和LLM辅助分析验证,提供了高可信度的重入漏洞标注。RSD则包含143个精心设计的最小合约示例,覆盖了Solidity 0.8+的现代特性,旨在通过语义边界案例测试检测工具的鲁棒性。这些数据集通过手动验证和场景化构造,解决了传统工具标签噪声和过时模式的问题,主要用于评估静态分析工具、机器学习模型及大语言模型在智能合约重入漏洞检测中的可靠性和适应性。
This study constructs two complementary datasets: Aggregated Benchmark and Reentrancy Scenarios Dataset (RSD). The Aggregated Benchmark, developed by institutions including the University of Venice, incorporates 432 real smart contracts sourced from prior academic datasets. It has been examined by multiple reviewers and validated via LLM-aided analysis, providing high-confidence reentrancy vulnerability annotations. The RSD, by contrast, contains 143 meticulously designed minimal contract examples covering modern features of Solidity 0.8+, and is intended to test the robustness of detection tools through semantic boundary case evaluations. Both datasets are built via manual validation and scenario-based construction, addressing the issues of label noise and outdated patterns in traditional tools. They are primarily used to evaluate the reliability and adaptability of static analysis tools, machine learning models, and large language models (LLMs) in reentrancy vulnerability detection for smart contracts.




