AI superposition, polysemanticity and mechanistic interpretability is fascinating. we have a chance of seeing what artificial neural networks are actually "thinking" using autoencoders to extract monosemantic features from polysemantic neurons.
Using these techniques we might be able to detect if AIs are being desceptive by peering into their brains, which will be useful if they try to enslave and/or kill us.
These terms probably makes no sense if you've never heard of them, I definitely didn't, but chris olah explains it well. Highly recommend the lex fridman podcast with him and other anthropic employees. if you have a spare... 5 hours.
https://podcasts.apple.com/ca/podcast/lex-fridman-podcast/id1434243584?i=1000676542285
I have no clue what you're talking about but it sounds very interesting .. seems like I need a day of vacation π
Well deserved, I enjoy your content π keep it up
Dear Santa .. #music #tunestr
#music #tunestr #nocomment https://youtu.be/QEphsKl1g9M
SLAY YOUR HEROES
I now am under sleep order. Wanna leave you with a thoughty tho π GN

Uhm yeah me too -.- the good news I can give you: the world will not end in the next hour. I'm still holding this station π
Fuck π
8:08
Listen to Untitled 5 by Unit Vector on #SoundCloud
https://on.soundcloud.com/j4JhQ
Try that. Quite heavy though!!
Nice! Like the "soundscape", as I call it in my head haha
This is why I am using Bandcamp π I want to give my money to artists directly and not to big record corpo or "license holders" or whatever maggots in the "value chain". This is all BS. Not sure about the artist terms there at the moment tbh, but afaik they still are good. At least I haven't heard anything else.
I posted this before but I listen to it daily so I'm just gonna post it again. No algo. What you gonna do about it? HA! #music


