AnchorSteer: Self-Discovered Concept Injection for Structure-Preserving Music Editing

ArXi:2605.31053v1 Announce Type: cross Controllable music editing is to modify high-level attributes while strictly preserving rhythmic and melodic structures. However, this task is challenged by a semantic-structural entanglement: steering methods often degrade structure to achieve editing performance, while structural adaptors suppress semantic responsiveness. We propose AnchorSteer, a framework that disentangles this tension by coupling structural anchoring with self-discovered semantic steering.