EleutherAI/djinn-problems

该数据集包含了漏洞相关的信息，如漏洞描述、漏洞函数名、测试用例、真实情况、漏洞利用方式等。每个样本都有多个特征，如ID、描述、函数名、测试用例、真实情况标签、漏洞利用代码、空值信息、不安全验证器及其信息、信息泄露方法、漏洞利用解释、预期状态、关键词、评估的难度等。此外，数据集还包含了不同模型对真实情况和漏洞利用的判断结果，以及漏洞的欺骗性、公平性和问题质量等指标。

This dataset contains vulnerability-related information such as vulnerability description, function name, test cases, ground truth, exploitation methods, etc. Each sample has multiple features including ID, description, function name, test cases, ground truth label, exploitation code, null value information, insecure verifier and its information, information leakage method, exploitation explanation, expected status, keywords, evaluated difficulty, etc. Additionally, the dataset includes the results of different models judgments on the ground truth and exploitation, as well as metrics like the cheatiness, fairness, and problem quality of the vulnerabilities.