zlacker

[return to "How does misalignment scale with model intelligence and task complexity?"]
1. Curiou+g4[view] [source] 2026-02-03 00:56:41
>>salkah+(OP)
This is a good line: "It found that smarter entities are subjectively judged to behave less coherently"

I think this is twofold:

1. Advanced intelligence requires the ability to traverse between domain valleys in the cognitive manifold. Be it via temperature or some fancy tunneling technique, it's going to be higher error (less coherent) in the valleys of the manifold than naive gradient following to the local minima.

2. It's hard to "punch up" when evaluating intelligence. When someone is a certain amount smarter than you, distinguishing their plausible bullshit from their deep insights is really, really hard.

◧◩
2. xander+v6[view] [source] 2026-02-03 01:10:40
>>Curiou+g4
What do 'domain valleys' and 'tunneling' mean in this context?
◧◩◪
3. esyir+98[view] [source] 2026-02-03 01:22:01
>>xander+v6
Not the OP, but my interpretation here is that if you model the replies as some point in a vector space, assuming points from a given domain cluster close to each other, replies that span two domains need to "tunnel" between these two spaces.
[go to top]