Position: Capability Control Should be a Separate Goal From Alignment

• URL: http://arxiv.org/abs/2602.05164v1
• Date: Thu, 05 Feb 2026 00:30:59 GMT
• Title: Position: Capability Control Should be a Separate Goal From Alignment
• Abstract summary: This paper argues that capability control should be treated as a distinct goal from alignment.<n>We organize capability control mechanisms across the model lifecycle into three layers.<n>Key challenges include the dual-use nature of knowledge and compositional generalization.
• Score: 43.64120192267
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Foundation models are trained on broad data distributions, yielding generalist capabilities that enable many downstream applications but also expand the space of potential misuse and failures. This position paper argues that capability control -- imposing restrictions on permissible model behavior -- should be treated as a distinct goal from alignment. While alignment is often context and preference-driven, capability control aims to impose hard operational limits on permissible behaviors, including under adversarial elicitation. We organize capability control mechanisms across the model lifecycle into three layers: (i) data-based control of the training distribution, (ii) learning-based control via weight- or representation-level interventions, and (iii) system-based control via post-deployment guardrails over inputs, outputs, and actions. Because each layer has characteristic failure modes when used in isolation, we advocate for a defense-in-depth approach that composes complementary controls across the full stack. We further outline key open challenges in achieving such control, including the dual-use nature of knowledge and compositional generalization.

Related papers

• Distributed Secure Learning Control for Large-scale Multirobots under Stealthy Actuator Attacks [24.92429415447018]
  This article proposes a distributed secure learning control framework for large-scale robots under stealthy attacks.<n>It employs a distributed attacker-actor-critic architecture to learn the optimal defense and attack policies online.<n>It is validated through comprehensive simulations and real-world experiments in multiple wheeled robots via various control tasks.
  arXiv  Detail & Related papers  (2026-09-07T00:39:28Z)
• Steering Language Models Before They Speak: Logit-Level Interventions [9.055997973281919]
  We propose a training-free inference-time logit intervention for controllable generation.<n>Our results show that statistically grounded logit steering can achieve large, consistent, and multi-task control gains.
  arXiv  Detail & Related papers  (2026-01-16T03:00:33Z)
• Plasma Shape Control via Zero-shot Generative Reinforcement Learning [17.3934551430283]
  We develop a novel framework for developing a versatile, zero-shot control policy from a large-scale offline dataset of PID-controlled discharges.<n>The resulting foundation policy can be deployed for diverse trajectory tracking tasks in a zero-shot manner without any task-specific fine-tuning.
  arXiv  Detail & Related papers  (2025-10-20T13:34:51Z)
• COMBO-Grasp: Learning Constraint-Based Manipulation for Bimanual Occluded Grasping [56.907940167333656]
  Occluded robot grasping is where the desired grasp poses are kinematically infeasible due to environmental constraints such as surface collisions.<n>Traditional robot manipulation approaches struggle with the complexity of non-prehensile or bimanual strategies commonly used by humans.<n>We introduce Constraint-based Manipulation for Bimanual Occluded Grasping (COMBO-Grasp), a learning-based approach which leverages two coordinated policies.
  arXiv  Detail & Related papers  (2025-02-12T01:31:01Z)
• Diffusion Predictive Control with Constraints [51.91057765703533]
  Diffusion predictive control with constraints (DPCC) is an algorithm for diffusion-based control with explicit state and action constraints.<n>We show through simulations of a robot manipulator that DPCC outperforms existing methods in satisfying novel test-time constraints.
  arXiv  Detail & Related papers  (2024-12-12T15:10:22Z)
• In-Distribution Barrier Functions: Self-Supervised Policy Filters that Avoid Out-of-Distribution States [84.24300005271185]
  We propose a control filter that wraps any reference policy and effectively encourages the system to stay in-distribution with respect to offline-collected safe demonstrations. Our method is effective for two different visuomotor control tasks in simulation environments, including both top-down and egocentric view settings.
  arXiv  Detail & Related papers  (2023-01-27T22:28:19Z)
• Lyapunov Density Models: Constraining Distribution Shift in Learning-Based Control [64.61499213110334]
  We seek a mechanism to constrain the agent to states and actions that resemble those that it was trained on. In control theory, Lyapunov stability and control-invariant sets allow us to make guarantees about controllers. density models allow us to estimate the training data distribution.
  arXiv  Detail & Related papers  (2022-06-21T16:49:09Z)
• Sparsity in Partially Controllable Linear Systems [56.142264865866636]
  We study partially controllable linear dynamical systems specified by an underlying sparsity pattern. Our results characterize those state variables which are irrelevant for optimal control.
  arXiv  Detail & Related papers  (2021-10-12T16:41:47Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.