points by fy20 1 day ago | 0 comments
There was a paper that popped up a month or two ago (I think published earlier though) about being able to inject "thoughts" into the neurons. E.g. Ask it what I should eat, and steer it towards "bread".
[–] krackers 22 hours ago
[–] noddybear 18 hours ago
This stuff (activation steering) has been around for a few years at this point
[–] square_usual 17 hours ago
Yeah, Golden Gate claude was in 2024! But I think the most recent/relevant thing was the monosemanticity discovery.
[–] tough 21 hours ago
mcnamara / goodfire couple weeks ago
https://www.anthropic.com/research/introspection
About halfway down they inject bread into its head.