Where Concept Erasure Should Occur: Concept-Layer Alignment in Text-to-Video Diffusion Models
For practitioners needing to remove sensitive concepts from text-to-video models, this work provides a principled method to improve erasure effectiveness without degrading output quality.
The paper identifies a representational bottleneck in text-to-video diffusion models, termed concept-layer topological alignment, where target concepts are more separable at certain depths. They propose CLEAR, a separability-driven optimization framework that enforces concept-layer alignment, achieving more precise concept suppression while preserving generative quality.
Text-to-video diffusion transformers encode semantic information unevenly across model depth, which constrains effective concept erasure. We identify a representational bottleneck, termed concept-layer topological alignment, under which target concepts exhibit higher separability at certain representational depths. Outside these depths, concept and non-target signals remain strongly entangled, limiting the effectiveness of depth-specific erasure. This observation reframes concept erasure as the problem of identifying representational depths where concept-non-target separation naturally emerges. Motivated by this structural constraint, we introduce CLEAR, a separability-driven optimization framework for concept erasure that explicitly enforces concept-layer alignment. CLEAR operationalizes this principle by formulating layer selection as an optimization problem over concept-non-target separability, rather than relying on layer-agnostic or heuristic choices. To enable this, we introduce a separability-aware objective that favors layers exhibiting stronger concept-non-target separation. Experiments on large-scale text-to-video diffusion models demonstrate that enforcing concept--layer alignment leads to more precise concept suppression while preserving overall generative quality.