Author

Marco Armenta

1 paper indexed here

We haven’t gathered this author’s papers yet. Follow them and we’ll fetch their work.

Not the right person? Other researchers publish under this name.

#machine learning Preprint Sep 2026

Hidden Activations are not Enough I: Knowledge Matrices as Higher Representations

The knowledge matrix of a trained feedforward network as a higher representation of its inputs is studied, grounding it in attribution theory and computing it by $C$ vector-Jacobian products instead of probing.

Marco Armenta · 0 citations