Tuned Lens

anchor

Findings

• IC-297Applying tuned-lens transformations to the residual stream decreases the apparent multi-layer SAE feature activity from 54–88% to 37–41% of total variance [primary]
• IC-561A linear direction in the residual stream at layer 16 of Llama3-8B-Instruct is causally necessary and sufficient for self-authorship claims: steering with it achieves 100% control over authorship assertions, and projecting it out reduces claims by 50-60%. [validation]
• IC-563The self-recognition vector's activation in Llama3-8B-Instruct is organized across depth: early layers (4-6) show diffuse perceptual activation to self-written text (present in both chat and base models), while layers 14-16 show a sharp decision-related peak at the output token that is present only in the chat model with role tags. [primary]