Activation Steering

Follow this topic Ask about this topic

Activation steering studies how edits to a model’s hidden activations can control its behavior at inference time, enabling changes without retraining the full model. Common approaches add or remove contrastive steering vectors, target sparse autoencoder features, neurons or attention heads, and use conditional or feedback-based controllers to adapt interventions to the input or generation state. Researchers apply these methods to language tasks such as reasoning, refusal, factuality and hallucination mitigation, as well as image and video generation, speech, and embodied agents; evaluations include benchmarks such as GSM8K, TruthfulQA, POPE and LIBERO. A central challenge is that readable or decodable features are not necessarily controllable, and effective edits can vary by context and architecture or cause collateral changes to fluency, utility and safety.

Latest papers

375 papers77 papers in the last four weeks, up 93% on the four weeks before. 0.5% of all new papers.

Parametric Social Identity Injection and Diversification in Public Opinion Simulation
Mar 17, 2026Hexi Wang, Yujia Zhou, Bangde Du +2Large Language Model-Based Social Simulation Representation Geometry in Language Models
DCST, Tsinghua University Beijing, China · Quancheng Laboratory Jinan, Shandong, China
Efficient Reasoning with Balanced Thinking
Mar 12, 2026Yulin Li, Tengyao Tu, Li Ding +5Efficient Language Model Reasoning Large Reasoning Models
Harbin Institute of Technology (Shenzhen) · Zhongguancun Academy · Huawei Noah’s Ark Lab +3
Global Evolutionary Steering: Refining Activation Steering Control via Cross-Layer Consistency
Mar 12, 2026Xinyan Jiang, Wenjing Yu, Di Wang +1Language Model Steering Activation Steering
Mohamed bin Zayed University of Artificial Intelligence (MBZUAI) · Provable Responsible AI and Data Analytics (PRADA) Lab · King Abdullah University of Science and Technology +3
Security-by-Design for LLM-Based Code Generation: Leveraging Internal Representations for Concept-Driven Steering Mechanisms
Mar 11, 2026Maximilian Wendlinger, Daniel Kowatsch, Konstantin Böttinger +1Secure Code Generation Code Language Models
Cognitive Security Technologies Fraunhofer Institute for Applied and Integrated Security Garching near Munich, Germany
PRISM- $Δ$ : Differential Subspace Steering for Prompt Highlighting in Large Language Models
Mar 11, 2026Yuyao Ge, Shenghua Liu, Yiwei Wang +5Language Model Steering LLM Prompting
♣Institute of Computing Technology, Chinese Academy of Sciences · △University of California, Merced
Adaptive Activation Cancellation for Hallucination Mitigation in Large Language Models
Mar 10, 2026Eric Yocam, Varghese Vaidyan, Gurcan Comert +3LLM Hallucination Mitigation Activation Steering
The Beacom College of Computer and Cyber Sciences, Dakota State University, Madison, SD 57042, USA. · Department of Computational Data Science and Engineering, North Carolina A&T State University, Greensboro, NC 27411, USA. · Department of Computer Science and Software Engineering, California Polytechnic State University, San Luis Obispo, CA 93407, USA. +2
Amnesia: Adversarial Semantic Layer Specific Activation Steering in Large Language Models
Mar 10, 2026Ali Raza, Gurang Gupta, Nikolay Matyunin +1Adversarial Attacks on LLMs LLM Jailbreak Attacks
Honda Research Institute Europe Offenbach, Hesse, Germany · IIT KharagPur Kharagpur, India
Curveball Steering: The Right Direction To Steer Isn't Always Linear
Mar 10, 2026Shivam Raval, Hae Jin Song, Linlin Wu +4Language Model Steering Activation Steering
PolypSteer: Counterfactual Endoscopic Synthesis via Training-Free Activation Steering
Mar 7, 2026Trong-Thang Pham, Loc Nguyen, Anh Nguyen +2Counterfactual Image Generation Medical Image Synthesis
University of Arkansas, Fayetteville AR, USA · University of Liverpool, Liverpool, UK · University of Houston, Houston TX, USA
COLD-Steer: Steering Large Language Models via In-Context One-step Learning Dynamics
Mar 6, 2026Kartik Sharma, Rakshit S. TrivediLanguage Model Steering Inference-Time Language Model Alignment
Georgia Institute of Technology · Massachusetts Institute of Technology
Enhancing Instruction Following of LLMs via Activation Steering with Dynamic Rejection
Mar 6, 2026Minjae Kang, Jaehyung KimLanguage Model Steering Instruction Following
Yonsei University
Observing and Controlling Features in Vision-Language-Action Models
Mar 5, 2026Hugo Buurmeijer, Carmen Amo Alonso, Aiden Swann +1VLM Interpretability Vision-Language-Action Models
Dept. of Aeronautics and Astronautics, Stanford University · Dept. of Mechanical Engineering, Stanford University · NVIDIA Research
Stable and Steerable Sparse Autoencoders with Weight Regularization
Mar 4, 2026Piotr Jedryszek, Oliver M. CrookSparse Autoencoders LLM Interpretability
Department of Biology, University of Oxford, Oxford, UK · Kavli Institute for Nanoscience Discovery, University of Oxford, Oxford, UK · Department of Chemistry, University of Oxford, Oxford, UK
Conditioned Activation Transport for T2I Safety Steering
Mar 3, 2026Maciej Chrabąszcz, Aleksander Szymczyk, Jan Dubiński +3Safe T2I Generation Activation Steering
NASK National Research Institute, Warsaw, Poland · Warsaw University of Technology, Warsaw, Poland · Tooplox +2
DreamReader: An Interpretability Toolkit for Text-to-Image Models
Mar 2, 2026Nirmalendu Prakash, Narmeen Oozeer, Michael Lan +6Diffusion Models T2I Diffusion Models
Singapore University of Technology and Design · Martian · Thoughtworks +2
Weight Updates as Activation Shifts: A Principled Framework for Steering
Feb 28, 2026Dyah Adila, John Cooper, Alexander Yun +2Activation Steering Parameter-Efficient Fine-Tuning
Department of Computer Science, University of Wisconsin-Madison
Steering and Rectifying Latent Representation Manifolds in Frozen Multi-modal LLMs for Video Anomaly Detection
Feb 27, 2026Zhaolin Cai, Fan Li, Huiyu Duan +2Video Anomaly Detection Multimodal Anomaly Detection
Xinjiang University · Xi’an Jiao Tong University · Shanghai Jiao Tong University +1
Suppressing Prior-Comparison Hallucinations in Radiology Report Generation via Semantically Decoupled Latent Steering
Feb 27, 2026Ao Li, Rui Liu, Mingjie Li +6Radiology Report Generation VLM Hallucination Mitigation
University of New South Wales · Australian Artificial Intelligence Institute, University of Technology Sydney · Stanford University +3
Interpreting and Steering State-Space Models via Activation Subspace Bottlenecks
Feb 26, 2026Vamshi Sunku Mohan, Kaustubh Gupta, Aneesha Das +1State Space Language Models State Space Models
AppViewX, Bangalore · Independent Researcher · Microsoft Research, Redmond
Facet-Level Persona Control by Trait-Activated Routing with Contrastive SAE for Role-Playing LLMs
Feb 22, 2026Wenqiu Tang, Zhen Wan, Takahiro Komamizu +1Persona-Conditioned Generation Large Language Model-Based Role-Play Simulation
Nagoya University, Nagoya, Aichi, Japan · Kyoto University, Kyoto, Japan
Post-Routing Arithmetic in Llama-3: Last-Token Result Writing and Rotation-Structured Digit Directions
Feb 22, 2026Yao YanNumerical Reasoning in Language Models Circuits in Language Models
Chongqing Normal University Chongqing Key Lab of Cognitive Intelligence and Intelligent Finance Chongqing, China
Understanding Unreliability of Steering Vectors in Language Models: Geometric Predictors and the Limits of Linear Approximations
Feb 19, 2026Joschka BraunLanguage Model Steering Activation Steering
Eberhard Karls University of Tübingen Faculty of Science
ODESteer: A Unified ODE-Based Steering Framework for LLM Alignment
Feb 19, 2026Hongjue Zhao, Haosen Sun, Jiangtao Kong +8Language Model Steering Activation Steering
University of Illinois Urbana-Champaign · Northwestern University · William & Mary +2
Discovering Universal Activation Directions for PII Leakage in Language Models
Feb 19, 2026Leo Marchyok, Zachary Coalson, Sungho Keum +2Privacy Leakage in Language Models LLM Interpretability
Oregon State University, Corvallis OR, USA · Korea Advanced Institute of Science & Technology, Daejeon, South Korea.
PERSONA: Dynamic and Compositional Inference-Time Personality Control via Activation Vector Algebra
Feb 17, 2026Xiachong Feng, Liang Zhao, Weihong Zhong +5Controllable Text Generation Personality Modeling in Language Models
Harbin Institute of Technology · The University of Hong Kong
Watch the Model Think: On-Policy Extraction of Activation Steering Vectors
Feb 15, 2026Xuanbo Su, Yingfang Zhang, Hao Luo +3Language Model Steering Representation Steering
Bairong Inc., Beijing, China · School of Mathematics, Harbin Institute of Technology, Harbin, China
Sparse Autoencoders are Capable LLM Jailbreak Mitigators
Feb 12, 2026Yannick Assogba, Jacopo Cortellazzi, Javier Abad +3Sparse Autoencoders LLM Jailbreak Attacks
1Apple · 2ETH Zurich
TADA! Tuning Audio Diffusion Models through Activation Steering
Feb 12, 2026Łukasz Staniszewski, Katarzyna Zaleska, Mateusz Modrzejewski +1Controllable Music Generation Audio Diffusion Models
Warsaw University of Technology · IDEAS Research Institute
General and Efficient Steering of Diffusion Models
Feb 11, 2026Qingsong Wang, Mikhail Belkin, Yusu WangDiffusion Model Guidance Activation Steering
Halıcıoğlu Data Science Institute, University of California San Diego, La Jolla, CA, USA
Mitigating Memorization Where It Happens
Feb 9, 2026Xuanqi Zhang, Haoyang Shang, Xiaoxiao LiMemorization in Language Models Privacy Leakage in Language Models
University of British Columbia · Vector Institute · Independent Researcher

Mar 17, 2026J/K move · Enter open · S save

Parametric Social Identity Injection and Diversification in Public Opinion Simulation

Hexi Wang, Yujia Zhou, Bangde Du, Qingyao Ai+1

DCST, Tsinghua University Beijing, China · Quancheng Laboratory Jinan, Shandong, China

Large Language Model-Based Social Simulation Representation Geometry in Language Models Activation Steering

Open paper AskPDF arXiv

Large language models (LLMs) have recently been adopted as synthetic agents for public opinion simulation, offering a promising alternative to costly and slow human surveys. Despite their scalability, current LLM-based simulation methods fail to capture social diversity, producing flattened inter-group differences and overly homogeneous responses across demographic groups. We identify this limitation as a Diversity Collapse phenomenon in LLM hidden representations, where distinct social identities become increasingly indistinguishable across layers. Motivated by this observation, we propose Parametric Social Identity Injection (PSII), a general framework that injects explicit, parametric representations of demographic attributes and value orientations directly into intermediate hidden states of LLMs. Unlike prompt-based persona conditioning, PSII enables fine-grained and controllable identity modulation at the representation level. Extensive experiments on the World Values Survey using multiple open-source LLMs show that PSII significantly improves distributional fidelity and diversity, reducing KL divergence to real-world survey data while enhancing overall diversity. This work provides new insights into representation-level control of LLM agents and advances scalable, diversity-aware public opinion simulation.