Viewing profile — josephmiller

In classic hacker news style, all the comments so far are nitpicking the definition of the word red-teaming. Yes, they are "red-teaming" their own research.

This is a terrifying alarm bell for humanity. Will we actually do anything about it or continue to drive faster and faster with our blindfold firmly on?

This seems like a tiny amount given the importantance of Chrome. Surely it would be rational for Google to pay 10x higher bounties?

Author here! This is correct. We did almost all this work on a Macbook Pro. Although for the pile-10k dataset analysis we used an A100 GPU because it would take many hours to run t…

Author here! I think this is reasonable but I have two responses. 1. It's kinda interesting because this is a clear case where the model must be thinking beyond the next token, whe…