It certainly wasn’t because the company is owned by a far-right South African billionaire at the same moment that the Trump admin is entertaining a plan to grant refugee status to white Afrikaners. /s
My partner is a real refugee. She was jailed for advocating democracy in her home country. She would have received a lengthy prison sentence after trial had she not escaped. This crap is bullshit. Btw, did you hear about the white-genocide happening in the USA? Sorry, I must have used Grok to write this. Go Elon! Cybertrucks are cool! Twitter isn’t a racist hellscape!
The stuff at the end was sarcasm, you dolt. Shut up.
While true, it doesn’t keep you safe from sleeper agent attacks.
These can essentially allow the creator of your model to inject (seamlessly, undetectably until the desired response is triggered) behaviors into a model that will only trigger when given a specific prompt, or when a certain condition is met. (such as a date in time having passed)
https://arxiv.org/pdf/2401.05566
It’s obviously not as likely as a company simply tweaking their models when they feel like it, and it prevents them from changing anything on the fly after the training is complete and the model is distributed, (although I could see a model designed to pull from the internet being given a vulnerability where it queries a specific URL on the company’s servers that can then be updated with any given additional payload) but I personally think we’ll see vulnerabilities like this become evident over time, as I have no doubts it will become a target, especially for nation state actors, to simply slip some faulty data into training datasets or fine-tuning processes that get picked up by many models.