As I said - The Scientific Python Scipy USA community has been goodd to me :)
They asked me to put the poster as a a reaseach paper. I think AST is a valuable thing if I am being honest. The Approach: Activation Sensory Transduction (AST)
Instead of converting activations into text for mech interp through autoencoding replacement models or NLA, AST routes a compressed version of the model's activation signals directly to a human operator's sensory system — bypassing language entirely.
The analogy is a radiologist reading an MRI scan. They don't need to transcribe the scan into words to act on what they see. A trained operator, given enough exposure to activation patterns, may develop the same kind of pattern-recognition intuition — without needing every internal state to be labeled in text first.
Some detail about the paper on my model recon site : model recon paper description
Actual research paper is here AST - Feel the model - reasearch paper
Big Thank you to the organizers and ans selectors at Scientific Python Scipy USA
This image explain a bit of more it: