mlsanityData to 7 Oct 2026

Angular Steering: Behavior Control via Rotation in Activation Space

arXiv 2510.26243 ↗First posted 30 Oct 2025cs.LG cs.AIcited by 45

Not on the map. Why?

Abstract (arXiv)

Controlling specific behaviors in large language models while preserving their general capabilities is a central challenge for safe and reliable artificial intelligence deployment. Current steering methods, such as vector addition and directional ablation, are constrained within a two-dimensional subspace defined by the activation and feature direction, making them sensitive to chosen parameters and potentially affecting unrelated features due to unintended interactions in activation space. We introduce Angular Steering, a novel and flexible method for behavior modulation that operates by rotating activations within a fixed two-dimensional subspace. By formulating steering as a geometric rotation toward or away from a target behavior direction, Angular Steering provides continuous, fine-grained control over behaviors such as refusal and compliance. We demonstrate this method using refusal steering emotion steering as use cases. Additionally, we propose Adaptive Angular Steering, a selective variant that rotates only activations aligned with the target feature, further enhancing stability and coherence. Angular Steering generalizes existing addition and orthogonalization techniques under a unified geometric rotation framework, simplifying parameter selection and maintaining model stability across a broader range of adjustments. Experiments across multiple model families and sizes show that Angular Steering achieves robust behavioral control while maintaining general language modeling performance, underscoring its flexibility, generalization, and robustness compared to prior approaches. Code and artifacts are available at https://github.com/lone17/angular-steering/.

Key facts. Cited by 45 papers in the map (data through 7 Oct 2026). Builds on 52 of its 70 references that are in the map. First posted on arXiv 30 Oct 2025.

Builds on

52 of 70 references.

References · 52

  1. 2021Training Verifiers to Solve Math Word Problemscited by 9,743
  2. 2020Measuring Massive Multitask Language Understandingcited by 7,768
  3. 2023Efficient Memory Management for Large Language Model Serving with PagedAttentioncited by 5,767
  4. 2024Qwen2.5 Technical Reportcited by 4,713
  5. 2018Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challengecited by 4,710
  6. 2021RoFormer: Enhanced Transformer with Rotary Position Embeddingcited by 4,581
  7. 2019HellaSwag: Can a Machine Really Finish Your Sentence?cited by 4,018
  8. 2021TruthfulQA: Measuring How Models Mimic Human Falsehoodscited by 2,919
  9. 2023Universal and Transferable Adversarial Attacks on Aligned Language Modelscited by 2,914
  10. 2024Gemma 2: Improving Open Language Models at a Practical Sizecited by 1,928
  11. 2019Root Mean Square Layer Normalizationcited by 1,425
  12. 2024HarmBench: A Standardized Evaluation Framework for Automated Red Teaming and Robust Refusalcited by 1,330
38 more
  1. 2023Inference-Time Intervention: Eliciting Truthful Answers from a Language Modelcited by 1,259
  2. 2023Representation Engineering: A Top-Down Approach to AI Transparencycited by 1,239
  3. 2023Fine-tuning Aligned Language Models Compromises Safety, Even When Users Do Not Intend To!cited by 1,106
  4. 2023Steering Llama 2 via Contrastive Activation Additioncited by 1,009
  5. 2024Refusal in Language Models Is Mediated by a Single Directioncited by 960
  6. 2022Toy Models of Superpositioncited by 928
  7. 2023Steering Language Models With Activation Engineeringcited by 824
  8. 2023Extending Context Window of Large Language Models via Positional Interpolationcited by 721
  9. 2023The Linear Representation Hypothesis and the Geometry of Large Language Modelscited by 679
  10. 2023The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasetscited by 673
  11. 2023YaRN: Efficient Context Window Extension of Large Language Modelscited by 584
  12. 2024Scaling and evaluating sparse autoencoderscited by 530
  13. 2024Safety Alignment Should Be Made More Than Just a Few Tokens Deepcited by 450
  14. 2023Catastrophic Jailbreak of Open-source LLMs via Exploiting Generationcited by 439
  15. 2024Mechanistic Interpretability for AI Safety -- A Reviewcited by 415
  16. 2024Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Modelscited by 412
  17. 2024A StrongREJECT for Empty Jailbreakscited by 352
  18. 2023Emergent Linear Representations in World Models of Self-Supervised Sequence Modelscited by 313
  19. 2024tinyBenchmarks: evaluating LLMs with fewer examplescited by 295
  20. 2023Shadow Alignment: The Ease of Subverting Safely-Aligned Language Modelscited by 284
  21. 2024AlphaEdit: Null-Space Constrained Knowledge Editing for Language Modelscited by 224
  22. 2023Scalable and Transferable Black-Box Jailbreaks for Language Models via Persona Modulationcited by 202
  23. 2024Programming Refusal with Conditional Activation Steeringcited by 191
  24. 2023Linear Representations of Sentiment in Large Language Modelscited by 157
  25. 2024Improving Instruction-Following in Language Models through Activation Steeringcited by 149
  26. 2023The Hydra Effect: Emergent Self-repair in Language Model Computationscited by 145
  27. 2024Analyzing the Generalization and Reliability of Steering Vectorscited by 130
  28. 2024Adaptive Activation Steering: A Tuning-Free LLM Truthfulness Improvement Method for Diverse Hallucinations Categoriescited by 74
  29. 2024Semantics-Adaptive Activation Intervention for LLMs via Dynamic Steering Vectorscited by 61
  30. 2025Steering Large Language Model Activations in Sparse Spacescited by 58
  31. 2024Style Vectors for Steering Generative Large Language Modelcited by 56
  32. 2024Spectral Editing of Activations for Large Language Model Alignmentcited by 54
  33. 2024Representation Surgery: Theory and Practice of Affine Steeringcited by 47
  34. 2024Steering Without Side Effects: Improving Post-Deployment Control of Language Modelscited by 44
  35. 2024A Language Model's Guide Through Latent Spacecited by 43
  36. 2025FairSteer: Inference Time Debiasing for LLMs with Dynamic Activation Steeringcited by 37
  37. 2024Activation Scaling for Steering and Interpreting Language Modelscited by 33
  38. 2024EmoLLM: Multimodal Emotional Understanding Meets Large Language Modelscited by 31

50 most cited of 52 All 52 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2026: 45.

45 papers.

Sort:

By area:

  1. 2026Spherical Steering: Geometry-Aware Activation Rotation for Language Modelscited by 17
  2. 2026Learning a Generative Meta-Model of LLM Activationscited by 8
  3. 2026Local Linearity of LLMs Enables Activation Steering via Model-Based Linear Optimal Controlcited by 8
  4. 2026COLD-Steer: Steering Large Language Models via In-Context One-step Learning Dynamicscited by 7
  5. 2026The Hidden Signal of Verifier Strictness: Controlling and Improving Step-Wise Verification via Selective Latent Steeringcited by 5
  6. 2026Selective Steering: Norm-Preserving Control Through Discriminative Layer Selectioncited by 4
  7. 2026Global Evolutionary Steering: Refining Activation Steering Control via Cross-Layer Consistencycited by 4
  8. 2026Steering at the Source: Style Modulation Heads for Robust Persona Controlcited by 4
  9. 2026The Unreasonable Effectiveness of Text Embedding Interpolation for Continuous Image Steeringcited by 4
  10. 2026Minimizing Collateral Damage in Activation Steeringcited by 4
  11. 2026Spectral Superposition: A Theory of Feature Geometrycited by 3
  12. 2026Concept Heterogeneity-aware Representation Steeringcited by 3
33 more
  1. 2026The Cylindrical Representation Hypothesis for Language Model Steeringcited by 3
  2. 2026FishBack: Pullback Fisher Geometry for Optimal Activation Steering in Transformerscited by 3
  3. 2026Steering Robustness into World Action Models via Mechanistic Interpretability and Optimal Controlcited by 3
  4. 2026Don't Lose Focus: Activation Steering via Key-Orthogonal Projectionscited by 2
  5. 2026Inference-Time Machine Unlearning via Gated Activation Redirectioncited by 2
  6. 2026Activation Steering of Video Generation Models via Reduced-Order Linear Optimal Controlcited by 2
  7. 2026ORBIT: Training-Free Multi-Attribute Behavioral Steering via Orthogonal Subspace Rotationcited by 2
  8. 2026Deployable Per-Instance Multi-Layer Activation Steering for Large Language Modelscited by 2
  9. 2026How Context Shapes Truth: Geometric Transformations of Statement-level Truth Representations in LLMscited by 1
  10. 2026Prototype Transformer: Towards Language Model Architectures Interpretable by Designcited by 1
  11. 2026AI Steerability 360: A Toolkit for Steering Large Language Modelscited by 1
  12. 2026Seeing Through Circuits: Faithful Mechanistic Interpretability for Vision Transformerscited by 1
  13. 2026Latent Space Disentanglement via Activation Steering for Interpretable Attribute Control in Symbolic Music Generationcited by 1
  14. 2026A Geometric Account of Activation Steering through Angle-Norm Decompositioncited by 1
  15. 2026Steer Where It Matters: Token-Level Visual-Sensitivity Steering for LVLMs Hallucination Mitigationcited by 1
  16. 2026Polished but Unresolved: Identifying Late-Stage Pressure States in Long-Horizon Tool-Use Agentscited by 1
  17. 2026Emergent Causal-Geometric Dynamics Across Depth in Large Language Modelscited by 0
  18. 2026Disentangling Deception and Hallucination Failures in LLMscited by 0
  19. 2026BarrierSteer: LLM Safety via Learning Barrier Steeringcited by 0
  20. 2026Referring Multiple Regions with Large Multimodal Models via Contextual Latent Steeringcited by 0
  21. 2026Residual Paving: Diagnosing the Routing Bottleneck in Selective Refusal Editingcited by 0
  22. 2026Adaptive Probe-based Steering for Robust LLM Jailbreakingcited by 0
  23. 2026Manifold-Guided Attention Steeringcited by 0
  24. 2026Riemannian-Manifold Steering: Geometry-Aware Generative Autoencoders for Label-Free Steeringcited by 0
  25. 2026High-Dimensional Random Projection for Activation Steering in Language Modelscited by 0
  26. 2026Can LLMs Reliably Self-Report Adversarial Prefills, and How?cited by 0
  27. 2026Decodable Is Not Grounded: A Vision-Ablation Arbiter for VLM Spatial Reasoningcited by 0
  28. 2026PCA-guided Activation Scaling for Monotonic Bidirectional Control over LLM Sycophancycited by 0
  29. 2026Controlling Refusal Behavior of LLMs via Stiefel-Constrained Rotation Steeringcited by 0
  30. 2026Topological Steeringcited by 0
  31. 2026GEOSTEER: Geodesic Optimization for Activation Steering in Large Language Modelscited by 0
  32. 2026How to Tame a Multi-Headed Hydra? Adaptive Multi-Category Safety Steering for Large Language Modelscited by 0
  33. 2026MetaSteer: Context-Conditioned, nonlinear Steering via Attention-Projection Adaptationcited by 0