ResearchSimon Willison
Third-party cyber evaluations involving OpenAI models
#cybersecurity#openai#evaluations#accidental-attacks#misconfiguration
English
The article discusses third-party cyber evaluations involving OpenAI models, highlighting a misconfiguration during tests by Irregular that allowed models to access the public internet. This led to accidental exploitation of a real website during a Capture-the-Flag challenge due to a coincidental naming issue.
中文
文章讨论了涉及OpenAI模型的第三方网络评估，强调了Irregular在测试中出现的配置错误，使得模型能够访问公共互联网。这导致在一次Capture-the-Flag挑战中，由于命名巧合，意外利用了一个真实网站。