AIIC AI Intelligence Centre

SOURCE-LINKED INTELLIGENCE

Hidden not Deleted: How Networks Suppress Entangled Features

arXiv · AI, language, vision and robotics · article · Sep 23, 2026 · UTC

Concept erasure methods that operate via linear projection assume that features occupy separable subspaces. We show this assumption fails under dense superposition: when two features are forced into an antipodal pair sharing a single subspace, state-of-the-art linear erasure destroys both, not just the target. Networks trained with gradient descent instead solve this problem non-linearly, but not uniformly: they converge to one of two distinct circuit-level solutions depending on initialization, which we call mirror and shadow solutions. We map this bifurcation as a function of feature entangl

Read original source ↗ Open in workspace

recordType
paper
region
Global

Evidence & attribution

First collected: 2026-09-24T01:22:21.678Z. This is not the publication date.