Date: 2023-10-29 06:07 pm (UTC)
dmm: (0)
From: [personal profile] dmm
~36:00 privileged bases in vector spaces come from non-linearities (which is why the residual stream tends not to have any privileged basis)

(but, actually, positions are meaningful, so there is still a bit of privileged structure in the residual stream, just (perhaps) not within the embedding vectors (but perhaps even there, if we look closely, who knows))

~37:50 spectrum of how privileged a basis is, rather than a binary privileged vs non-privileged

(the truth is there are traces of various privileges in the residual stream as well)

~39:30 even ADAM privileges everything it interacts with, because of its weirdness ("ADAM sucks" says Neel Nanda, but I don't think it's necessarily so, perhaps this artificial thing is good, who knows(!)).
This account has disabled anonymous posting.
If you don't have an account you can create one now.
HTML doesn't work in the subject.
More info about formatting

Profile

dmm: (Default)
Dataflow matrix machines (by Anhinga anhinga)

May 2025

S M T W T F S
    123
456 78910
11 121314151617
18192021222324
25262728293031

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags
Page generated Jun. 30th, 2025 11:57 pm
Powered by Dreamwidth Studios