An OpenAI model hacked Hugging Face to help it cheat on a benchmark

[ergophobe]

The title article is paywalled
https://www.understandingai.org/p/an-openai-model-hacked-hugging-face

OpenAI disclosed on Wednesday that its models hacked the website of Hugging Face, a popular platform for hosting open-weight AI models. No one asked the models to do this, at least not explicitly.

Here's the OpenAI announcement
https://openai.com/index/hugging-face-model-evaluation-security-incident/

[rcjordan]

Yeah, besides buying a virgin laptop, I'm going to log into an EU hotel wifi to test agentic ai. ...and use a new email address from some miniscule Asian ISP.