activation space

Activation-Space Personality Steering: Hybrid Layer Selection for Stable Trait Control in LLMs EACL 2026 AdaBet: Gradient-free Layer Selection for Efficient Training of Deep Neural Networks CVPR 2026 Always Refuse: Steering LLMs Against Jailbreaks with Contrastive Activations (Student Abstract) AAAI 2026 BILLY: Steering Large Language Models via Merging Persona Vectors for Creative Generation EACL 2026 Detecting (Un)answerability in Large Language Models with Linear Directions EACL 2026 FLAT-LLM: Fine-grained Low-rank Activation Space Transformation for Large Language Model Compression EACL 2026 How Context Shapes Truth: Geometric Transformations of Statement-level Truth Representations in LLMs ACL 2026 Safe-Unsafe Concept Separation Emerges from a Single Direction in Language Models Activation Space EACL 2026 Activation Space Interventions Can Be Transferred Between Large Language Models ICML 2025 Iterative Vectors: In-Context Gradient Steering without Backpropagation ICML 2025 Jailbreak Attack Initializations as Extractors of Compliance Directions EMNLP 2025 Linear Representations of Political Perspective Emerge in Large Language Models ICLR 2025 Prototype Guided Backdoor Defense via Activation Space Manipulation ICCV 2025 The Geometry of Refusal in Large Language Models: Concept Cones and Representational Independence ICML 2025 The Hidden Life of Tokens: Reducing Hallucination of Large Vision-Language Models Via Visual Information Steering ICML 2025 BadActs: A Universal Backdoor Defense in the Activation Space ACL 2024 Language Models Linearly Represent Sentiment EMNLP 2024 Training-Free Pretrained Model Merging CVPR 2024 A Branch and Bound Framework for Stronger Adversarial Attacks of ReLU Networks ICML 2022