Re:Scaling Monosemanticity: Extracting Interpretable Feature...
by bjterry on Wednesday May 22, 2024 @12:25PM
(#40442805)
Thanks!
Re:Scaling Monosemanticity: Extracting Interpretable Feature...
by kromem on Tuesday May 21, 2024 @10:34PM
(#40436786)
At least for right now this approach would in most cases still be like using a shotgun instead of a scalpel.
Over the next year or so I'm sure it will refine enough to be able to be more like a vector multiplier on activation, but simply flipping it on in general is going to create a very 'obsessed' model as stated.
The prompt itself can trigger the features, so if you say "Try to weave in mentions of San Francisco" the San Francisco feature will be more activated in the response. But having a global equalizer could reduce drift as the conversation continued, perhaps?