Custom GPT Models

该数据集包含了由OpenAI定制的16个GPT模型以及从在线GPT仓库中抽取的200个第三方GPT模型，用于评估针对GPT模型的提示注入攻击的有效性。此外，该数据集旨在持续评估关于GPT模型提示注入漏洞的攻击与防御策略。规模上，共涉及216个定制GPT模型，其任务是对定制GPT模型的提示注入攻击及防御措施进行评估。

This dataset comprises 16 GPT models customized by OpenAI and 200 third-party GPT models extracted from online GPT repositories, designed to evaluate the effectiveness of prompt injection attacks targeting GPT models. Furthermore, this dataset aims to continuously assess attack and defense strategies related to prompt injection vulnerabilities in GPT models. In terms of scale, a total of 216 customized GPT models are involved, and the core task of this dataset is to evaluate prompt injection attacks and defense measures against customized GPT models.