Attention Manifolds: Steering or Blocking Language Models by Editing Learned B-Spline Surfaces

Abstract:In standard transformer attention, a source token sends the same value vector to every receiver. The query determines \emph{how much} to attend but not \emph{what} to extract. This work introduces \textbf{attention manifolds}: learned 2D B-spline surfaces $S_d(q_d, k_d)$ that modulate each value dimension based on the query-key interaction. Each surface is a tensor-product cubic B-spline initialized to zero, preserving pretrained behavior. Applied to LLaMA 3.2-1B-Instruct and 3B-Instruct, attention manifolds reduce WikiText-2 validation perplexity by 2--2.5 points with 0.3\% parameter overhead. Across 112 diverse prompts, surfaces change greedy-decoded output for 69\% (1B) to 83\% (3B) of cases, with the strongest effects on ambiguous and polysemous inputs (94--100\% change rate). The surfaces improve output quality: correcting factual errors (\emph{``the CAP theorem has three main components''} $\to$ \emph{``it is impossible to guarantee all three''}), increasing precision (\emph{``impossible to know certain properties''} $\to$ \emph{``impossible to know both position and momentum''}), and adding specificity (a generic quote $\to$ an attributed Saint Augustine citation, consistently at both scales). The learned surfaces are also mechanically editable: inverting a layer's coefficients changes greedy output for 9/10 prompts (KL~0.010), providing a geometric mechanism for model steering. Setting surface coefficients to $-1$ creates ``attention walls'' that block value flow through specific dimensions. In a preliminary experiment, a layer-wide wall redirects an explosive-device prompt from specific instructions to general educational content, suggesting a path toward safety-oriented manifold shaping.

Machine Learning (cs.LG)

arXiv:2610.00257 [cs.LG]

(or arXiv:2610.00257v1 [cs.LG] for this version)

https://doi.org/10.48550/arXiv.2610.00257 Focus to learn more arXiv-issued DOI via DataCite (pending registration)

arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.

Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.

Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.