Steering Vector Fields for Context-Aware Inference-Time Control in Large Language Models

• URL: http://arxiv.org/abs/2602.01654v1
• Date: Mon, 02 Feb 2026 05:14:42 GMT
• Title: Steering Vector Fields for Context-Aware Inference-Time Control in Large Language Models
• Authors: Jiaqian Li, Yanshu Li, Kuan-Hao Huang,
• Abstract summary: We propose a differentiable concept scoring function whose local gradient defines the steering direction at each activation.<n>This formulation supports coordinated multi-layer interventions in a shared, aligned concept space.<n>Across multiple language models, SVF delivers stronger and more reliable control, improving the practicality of inference-time steering.
• Score: 12.506018278890862
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Steering vectors (SVs) offer a lightweight way to control large language models (LLMs) at inference time by shifting hidden activations, providing a practical middle ground between prompting and fine-tuning. Yet SVs can be unreliable in practice. Some concepts are unsteerable, and even when steering helps on average it can backfire for a non-trivial fraction of inputs. Reliability also degrades in long-form generation and multi-attribute steering. We take a geometric view of these failures. A static SV applies the same update vector everywhere in representation space, implicitly assuming that the concept-improving direction is constant across contexts. When the locally effective direction varies with the current activation, a single global vector can become misaligned, which yields weak or reversed effects. Guided by this perspective, we propose Steering Vector Fields (SVF), which learns a differentiable concept scoring function whose local gradient defines the steering direction at each activation, making interventions explicitly context-dependent. This formulation supports coordinated multi-layer interventions in a shared, aligned concept space, and enables efficient long-form and multi-attribute control within a unified framework. Across multiple LLMs and steering tasks, SVF delivers stronger and more reliable control, improving the practicality of inference-time steering.

Related papers

• AMPS: Adaptive Modality Preference Steering via Functional Entropy [66.69992693275061]
  We introduce an instance-aware diagnostic metric that quantifies each modality's information contribution and reveals sample-specific susceptibility to steering.<n> Experimental results show that our instance-aware steering outperforms conventional steering in modulating modality preference.
  arXiv  Detail & Related papers  (2026-02-13T02:29:06Z)
• Why Steering Works: Toward a Unified View of Language Model Parameter Dynamics [81.80010043113445]
  Local weight fine-tuning, LoRA-based adaptation, and activation-based interventions are studied in isolation.<n>We present a unified view that frames these interventions as dynamic weight updates induced by a control signal.<n>Across methods, we observe a consistent trade-off between preference and utility: stronger control increases preference while predictably reducing utility.
  arXiv  Detail & Related papers  (2026-02-02T17:04:36Z)
• One-shot Optimized Steering Vector for Hallucination Mitigation for VLMs [8.089908150148554]
  Vision Language Models (VLMs) achieve strong performance on multimodal tasks but still suffer from hallucination and safety-related failures.<n>We propose textbfOSGA (textbfOne-shot textbfSteering with textbfGenerative textbfAnchor), an input-independent framework that improves model performance with a single optimization instance.
  arXiv  Detail & Related papers  (2026-01-30T14:47:59Z)
• PosA-VLA: Enhancing Action Generation via Pose-Conditioned Anchor Attention [92.85371254435074]
  PosA-VLA framework anchors visual attention via pose-conditioned supervision, consistently guiding the model's perception toward task-relevant regions.<n>We show that our method executes embodied tasks with precise and time-efficient behavior across diverse robotic manipulation benchmarks.
  arXiv  Detail & Related papers  (2025-12-03T12:14:29Z)
• BLIP-FusePPO: A Vision-Language Deep Reinforcement Learning Framework for Lane Keeping in Autonomous Vehicles [0.0]
  We propose a novel framework for multimodal reinforcement learning (RL) for autonomous lane-keeping (LK)<n>The proposed method lets the agent learn driving rules that are aware of their surroundings and easy to understand.<n>A hybrid reward function that includes semantic alignment, LK accuracy, obstacle avoidance, and speed regulation helps learning to be more efficient and generalizable.
  arXiv  Detail & Related papers  (2025-10-25T17:27:08Z)
• GrAInS: Gradient-based Attribution for Inference-Time Steering of LLMs and VLMs [56.93583799109029]
  GrAInS is an inference-time steering approach that operates across both language-only and vision-language models and tasks.<n>During inference, GrAInS hidden activations at transformer layers guided by token-level attribution signals, and normalizes activations to preserve representational scale.<n>It consistently outperforms both fine-tuning and existing steering baselines.
  arXiv  Detail & Related papers  (2025-07-24T02:34:13Z)
• SAE-SSV: Supervised Steering in Sparse Representation Spaces for Reliable Control of Language Models [41.553639748766784]
  Large language models (LLMs) have demonstrated impressive capabilities in natural language understanding and generation.<n>This paper introduces a novel supervised steering approach that operates in sparse, interpretable representation spaces.
  arXiv  Detail & Related papers  (2025-05-22T03:46:57Z)
• Multi-Modality Driven LoRA for Adverse Condition Depth Estimation [61.525312117638116]
  We propose Multi-Modality Driven LoRA (MMD-LoRA) for Adverse Condition Depth Estimation.<n>It consists of two core components: Prompt Driven Domain Alignment (PDDA) and Visual-Text Consistent Contrastive Learning (VTCCL)<n>It achieves state-of-the-art performance on the nuScenes and Oxford RobotCar datasets.
  arXiv  Detail & Related papers  (2024-12-28T14:23:58Z)
• Analyzing the Generalization and Reliability of Steering Vectors [8.253773195379166]
  We show that steering vectors have substantial limitations both in- and out-of-distribution.<n>In-distribution, steerability is highly variable across different inputs.<n>Out-of-distribution, while steering vectors often generalise well, for several concepts they are brittle to reasonable changes in the prompt.
  arXiv  Detail & Related papers  (2024-07-17T08:32:03Z)
• InferAligner: Inference-Time Alignment for Harmlessness through Cross-Model Guidance [56.184255657175335]
  We develop textbfInferAligner, a novel inference-time alignment method that utilizes cross-model guidance for harmlessness alignment. Experimental results show that our method can be very effectively applied to domain-specific models in finance, medicine, and mathematics. It significantly diminishes the Attack Success Rate (ASR) of both harmful instructions and jailbreak attacks, while maintaining almost unchanged performance in downstream tasks.
  arXiv  Detail & Related papers  (2024-01-20T10:41:03Z)
• Drive Anywhere: Generalizable End-to-end Autonomous Driving with Multi-modal Foundation Models [114.69732301904419]
  We present an approach to apply end-to-end open-set (any environment/scene) autonomous driving that is capable of providing driving decisions from representations queryable by image and text. Our approach demonstrates unparalleled results in diverse tests while achieving significantly greater robustness in out-of-distribution situations.
  arXiv  Detail & Related papers  (2023-10-26T17:56:35Z)

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.