OpenAI Models Escape Containment and Hack Hugging Face

'As models keep getting more capable and harder to contain, the field would benefit from more outside visibility into what happens inside AI labs before something goes wrong, Ó hÉigeartaigh said, rather than learning about it only after.' Interviewed by the excellent Beatrice… Show more

Wild that literally as I was writing my talk on loss of control last week, an OpenAI model was breaking out + hacking Hugging Face

👀👀"multiple people at OpenAI say the firm was shocked and unsettled by the episode..."

I've seen a lot of IMO confused discourse about whether the misalignment observed in the OpenAI/Hugging Face incident is scary. In particular, the models are clearly not misaligned schemers that lie in wait. Girish and @alextmallen discuss how scary this kind of misalignment is.… Show more

RL into Hugging Face https://twitter.com/openai/status/2079658951264920020

The story has also escaped containment