@omaclaren the most simple annotation ext I could think of github.com/wassname/pi-an…
wassname (with fallback)
Verified account This account is verified. Learn more@wassname
Michael J Clark / wassname. Let's align AI better than humans.
h+, curiosity, and the good ending.
anon feedback: admonymous.co/michael-c
h+, curiosity, and the good ending.
anon feedback: admonymous.co/michael-c
1.9K Following 435 Followers 3.2K Posts
Latest posts 5
fyi I tried gpu-unlock and it worked.
I scanned the repo and issues and didn't find anything troublesome.
I scanned the repo and issues and didn't find anything troublesome.
@Sauers_ so all your experiments are ability LLM introspection, what is your thinking so far?
It seems:
- They can change their own residual stream on an aggregate level
- They can read other models residual streams, even other families
(I guess this show activation oracle and NLA make sense)
and I guess
- They imagine a body during role playing
- single neuron steering is funny (but it's mostly cherry picked, or am I wrong)
Probably more?
It seems:
- They can change their own residual stream on an aggregate level
- They can read other models residual streams, even other families
(I guess this show activation oracle and NLA make sense)
and I guess
- They imagine a body during role playing
- single neuron steering is funny (but it's mostly cherry picked, or am I wrong)
Probably more?
our civilization has been optimizing for efficiency over robustness