LARC: Low-Rank Adaptive Residual Connections for Learning in Frozen Models

Excerpt

arXiv:2609.40063v1 Announce Type: new Abstract: Low-Rank Adaptive Residual Connections (LARC) give a frozen model a compact numerical state that can learn from feedback. The map $h+BAh$ adds a low-rank correction to a hidden representation. A slow state $\rho$ learns starting factors across tasks; a private fast state $\Phi$ copies them, changes with feedback, and resets to the trained initialization. This report specifies an input-side realization of the numerical policy carrier in Memory-Media