activation steering

ARGUS: Defending Against Multimodal Indirect Prompt Injection via Steering Instruction-Following Behavior CVPR 2026 Can Activation Steering Generalize Across Languages? A Study on Syllogistic Reasoning in Language Models EACL 2026 Causal Activation Steering via Sparse Mediation EACL 2026 CoFact: Dynamic Coordination of Attention Heads for Improving Factual Consistency in LLMs AAAI 2026 Competence Collapse in Code-Mixed Generation: Spectral Evidence and Mechanistic Recovery via Cross-Lingual Activation Steering EACL 2026 Compiling Activation Steering into Weights via Null-Space Constraints for Stealthy Backdoors ACL 2026 Continuous Interpretive Steering for Scalar Diversity ACL 2026 CoSToM: Causal-oriented Steering for Intrinsic Theory-of-Mind Alignment in Large Language Models ACL 2026 Detecting What Queries Seek: Steering LLM Safety with FFN Output Activation Monitoring ACL 2026 DFKI-MLT at SemEval-2026 TASK 7: Steering Multilingual Models Towards Cultural Knowledge SemEval 2026 Diagnosing and Repairing Unsafe Channels in Vision-Language Models via Causal Discovery and Dual-Modal Safety Subspace Projection CVPR 2026 Dictionary-Aligned Concept Control for Safeguarding Multimodal LLMs CVPR 2026 Differentiated Directional Intervention: A Framework for Evading LLM Safety Alignment AAAI 2026 d’Olle Grieze at SemEval-2026 Task 11: Comparing the Impact of Supervised Fine-Tuning and Activation Steering on Mitigating Content Effect Bias in Syllogistic Reasoning SemEval 2026 DSO: Direct Steering Optimization for Bias Mitigation CVPR 2026 Eliciting Chain-of-Thought in Base LLMs via Gradient-Based Representation Optimization AAAI 2026 Embracing Anisotropy: Turning Massive Activations into Interpretable Control Knobs for Large Language Models ACL 2026 From Weights to Activations: Is Steering the Next Frontier of Adaptation? ACL 2026 GrAInS: Gradient-based Attribution for Inference-Time Steering of LLMs and VLMs ACL 2026 Impatient Users Confuse AI Agents: High-fidelity Simulations of Human Traits for Testing Agents ACL 2026 Latent Agents: A Post-Training Procedure for Internalized Multi-Agent Debate ACL 2026 Mitigating Content Effects on Reasoning in Language Models Through Fine-Grained Activation Steering AAAI 2026 Principled Steering via Null-space Projection for Jailbreak Defense in Vision-Language Models CVPR 2026 PrivSV: Differentially Private Steering Vector for Large Language Models AAAI 2026 Probing Bias Formation in Medical LLMs through Activation Steering ACL 2026