Steerable but Not Decodable: Function Vectors Operate Beyond the Logit Lens
Paper • 2604.02608 • Published • 1
Suhail Nadaf
suhailnadaf509
AI & ML interests
Mechanistic interpretability
Recent Activity
authored a paper 2 days ago
Steerable but Not Decodable: Function Vectors Operate Beyond the Logit Lens upvoted a paper 3 days ago
Steerable but Not Decodable: Function Vectors Operate Beyond the Logit Lens liked a model 5 months ago
google/gemma-scope-2Organizations
None yet
authored a
paper 2 days ago