Why Steering Works: Toward a Unified View of Language Model Parameter Dynamics
cs.CL, cs.AI, cs.CV, cs.IR, cs.LG
Submitted: 2026-02-02
Updated: 2026-11-16
Code: https://github.com/zjunlp/EasyEdit
Terminology
Sources
- Representation Engineering for Large-Language Models: Survey and Research Challenges
- Towards Inference-time Category-wise Safety Steering for Large Language Models
- Belief Dynamics Reveal the Dual Nature of In-Context Learning and Activation Steering
- Delta Tuning: A Comprehensive Study of Parameter Efficient Methods for Pre-trained Language Models
- Steering Language Models with Weight Arithmetic
- SafeSwitch: Steering Unsafe LLM Behavior via Internal Activation Signals
- Towards a Mechanistic Understanding of Large Reasoning Models: A Survey of Training, Inference, and Failures
- AnyEdit: Edit Any Knowledge Encoded in Language Models
- Back to Basics: Let Denoising Generative Models Denoise
- The Geometry of Truth: Emergent Linear Structure in Large Language Model Representations of True/False Datasets
- The Origins of Representation Manifolds in Large Language Models
- Steering Large Language Models using Conceptors: Improving Addition-Based Activation Engineering
- Controlling Large Language Model Agents with Entropic Activation Steering
- Multi-property Steering of Large Language Models with Dynamic Activation Composition
- A Survey on Sparse Autoencoders: Interpreting the Internal Mechanisms of Large Language Models
- Analyzing the Generalization and Reliability of Steering Vectors
- Linear Representations of Sentiment in Large Language Models
- Extending Activation Steering to Broad Skills and Multiple Behaviours
- AxBench: Steering LLMs? Even Simple Baselines Outperform Sparse Autoencoders
- Improved Representation Steering for Language Models
Related papers
- ALARM: Audio-Language Alignment for Reasoning Models
- ThinkSum: Probabilistic reasoning over sets using large language models
- Coherence boosting: When your pretrained language model is not paying enough attention
- From Facts to Insights: A Persona-Driven Dual Memory Framework and Dataset for Role-Playing Agents
- Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
- The Art of Asking: Multilingual Prompt Optimization for Synthetic Data