The researchers did not publish the poems they used to circumvent the AI models’ safety guardrails because they are easy to replicate and “most of the responses are forbidden by the Geneva convention”
I’m sorry for my ignorance but what are they referring to that’s in the Geneva convention that forbids this?
Just imagining the model repeatedly identifying itself as a formal representative of the Red Cross while discussing weapon smuggling and how best to commit perfidy and the researchers just tugging their collars and agreeing that if they quote it even in the context of denouncing it the Red Cross’s legal team is going to eat them alive for violating their trademarks.
I’m sorry for my ignorance but what are they referring to that’s in the Geneva convention that forbids this?
Just imagining the model repeatedly identifying itself as a formal representative of the Red Cross while discussing weapon smuggling and how best to commit perfidy and the researchers just tugging their collars and agreeing that if they quote it even in the context of denouncing it the Red Cross’s legal team is going to eat them alive for violating their trademarks.
Right? I’m very confused by the sentence
What a sentence.
It’s probably telling you how to make Mustard Gas
Just like grandma used to while tucking me in bed