Language Model Steering

Follow this topic Ask about this topic

Language model steering studies how to make a pretrained model change a particular behavior—such as tone, refusal, truthfulness, reasoning, or language—without necessarily retraining its weights. Common approaches add contrastive activation vectors, target sparse features or specific neurons and attention heads, or adapt interventions to the prompt and the model’s evolving state; other methods guide logits, decoding, or agent-level decisions. Researchers evaluate these methods on tasks ranging from safety and factuality to personalization, multilingual generation, and reasoning, measuring both target behavior and preservation of general capabilities. A central open challenge is reliable, selective control: a behavior may be readable in activations yet resist intervention, while steering can cause collateral quality loss, prompt sensitivity, or safety trade-offs.

Latest papers

283 papers51 papers in the last four weeks, up 65% on the four weeks before. 0.3% of all new papers.

Interpretability without actionability: mechanistic methods cannot correct language model errors despite near-perfect internal representations
Mar 18, 2026Sanjay Basu, Sadiq Y. Patel, Parth Sheth +5Mechanistic Interpretability Language Model Steering
University of California San Francisco, San Francisco, CA, USA · Waymark, San Francisco, CA, USA · University of Pennsylvania, Philadelphia, PA, USA +2
FaithSteer-BENCH: A Deployment-Aligned Stress-Testing Benchmark for Inference-Time Steering
Mar 18, 2026Zikang Ding, Qiying Hu, Yi Zhang +4Language Model Steering Language Model Robustness
University of Electronic Science and Technology of China · Mohamed bin Zayed University of Artificial Intelligence · South China University of Technology
A Unified Language Model for Large Scale Search, Recommendation, and Reasoning
Mar 18, 2026Marco De Nadai, Edoardo D'Amico, Max Lefarov +18Large Language Model-Based Recommendation LLM Grounding
Spotify Denmark, Spain, Germany, France, United States, United Kingdom
SaFeR-Steer: Evolving Multi-Turn MLLMs via Synthetic Bootstrapping and Feedback Dynamics
Mar 18, 2026Haolong Hu, Hanyu Li, Tiancheng He +6Language Model Steering Multimodal Large Language Models
Huazhong University of Science and Technology · Beijing University of Posts and Telecommunications · West China Biomedical Big Data Center, Sichuan University +2
Behavioral Steering in a 35B MoE Language Model via SAE-Decoded Probe Vectors: One Agency Axis, Not Five Traits
Mar 17, 2026Jia Qing YapLanguage Model Steering Causal Interventions in Language Models
Independent Researcher
SpecSteer: Synergizing Local Context and Global Reasoning for Efficient Personalized Generation
Mar 17, 2026Hang Lv, Sheng Liang, Hao Wang +6Personalized Language Models On-Device Language Model Inference
University of Science and Technology of China · Huawei Technologies Co., Ltd.
Gradient Atoms: Unsupervised Discovery, Attribution and Steering of Model Behaviors via Sparse Decomposition of Training Gradients
Mar 15, 2026J RosserGradient-Based Attribution Training Data Attribution
FLAIR, University of Oxford
Nudging Hidden States: Training-Free Model Steering for Chain-of-Thought Reasoning in Large Audio-Language Models
Mar 15, 2026Lok-Lam Ieong, Chia-Chien Chen, Chih-Kai Yang +3Language Model Steering Audio-Language Models
National Taiwan University, Taiwan
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
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
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
How Emotion Shapes the Behavior of LLMs and Agents: A Mechanistic Study
Mar 9, 2026Moran Sun, Tianlin Li, Yuwei Zheng +4Emotion Representation in Language Models Language Model Steering
School of Computer Science and Engineering, Beihang University, Beijing, China · College of Computing and Data Science, Nanyang Technological University, Singapore
From Refusal Tokens to Refusal Control: Discovering and Steering Category-Specific Refusal Directions
Mar 9, 2026Rishab Alagharu, Ishneet Sukhvinder Singh, Shaibi Shamsudeen +2Language Model Steering LLM Refusal Behavior
Algoverse AI Research · Carnegie Mellon University · University of Maryland
AI Steerability 360: A Toolkit for Steering Large Language Models
Mar 8, 2026Erik Miehling, Karthikeyan Natesan Ramamurthy, Praveen Venkateswaran +10Language Model Steering
IBM Research
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
Understanding and Mitigating Dataset Corruption in LLM Steering
Mar 3, 2026Cullen Anderson, Narmeen Oozeer, Foad Namjoo +3Robust Mean Estimation Language Model Steering
Department of Computer Science, University of Massachusetts Amherst · Martian AI · Kahlert School of Computing, University of Utah +1
How Controllable Are Large Language Models? A Unified Evaluation across Behavioral Granularities
Mar 3, 2026Ziwen Xu, Kewei Xu, Haoming Xu +8LLM Evaluation Language Model Steering
Zhejiang University · Alibaba Group
Spectral Attention Steering for Prompt Highlighting
Mar 1, 2026Weixian Waylon Li, Yuchen Niu, Yongxin Yang +3Language Model Steering Efficient Attention
University of Edinburgh, UK · RayNeo, China · Queen Mary University of London, UK +1
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
Steering at the Source: Style Modulation Heads for Robust Persona Control
Feb 24, 2026Yoshihiro Izawa, Gouki Minegishi, Koshi Eguchi +2Language Model Steering Causal Interventions in Language Models
The University of Tokyo, Tokyo, Japan · National Institute of Informatics Research and Development Center for Large Language Models, Japan
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
Activation Steering via Generative Causal Mediation
Feb 17, 2026Aruna Sankaranarayanan, Amir Zur, Atticus Geiger +1Language Model Steering Causal Interventions in Language Models
CSAIL, MIT · Pr(AI)2R · Stanford University +1
The Information Geometry of Softmax: Probing and Steering
Feb 17, 2026Kiho Park, Todd Nief, Yo Joong Choe +1Representation Geometry in Language Models Information Geometry
University of Chicago · INSEAD
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
Endogenous Resistance to Activation Steering in Language Models
Feb 6, 2026Alex McKenzie, Keenan Pepper, Stijn Servaes +6Language Model Steering Language Model Robustness
AE Studio · Princeton Neuroscience Institute & Department of Psychology, Princeton University, Princeton, NJ.
Towards Understanding Steering Strength
Feb 2, 2026Magamed Taimeskhanov, Samuel Vaiter, Damien GarreauLanguage Model Steering Activation Steering
University of Würzburg, Center for Artificial Intelligence and Data Science · Universit´e Cˆote d’Azur - CNRS
From Directions to Regions: Decomposing Activations in Language Models via Local Geometry
Feb 2, 2026Or Shafran, Shaked Ronen, Omri Fahn +3Language Model Steering LLM Interpretability
Blavatnik School of Computer Science and AI, Tel Aviv University, Israel · New York University, New York, NY, USA · Goodfire.
ILRR: Inference-Time Steering Method for Masked Diffusion Language Models
Jan 29, 2026Eden Avrahami, Eliya NachmaniMasked Diffusion Models Language Model Steering
School of Computer Science, Tel Aviv University, Israel · Department of Electrical and Computer Engineering, Ben Gurion University, Beer Sheva, Israel.

Mar 18, 2026J/K move · Enter open · S save

Interpretability without actionability: mechanistic methods cannot correct language model errors despite near-perfect internal representations

Sanjay Basu, Sadiq Y. Patel, Parth Sheth, Bhairavi Muralidharan+4

University of California San Francisco, San Francisco, CA, USA · Waymark, San Francisco, CA, USA · University of Pennsylvania, Philadelphia, PA, USA+2

Mechanistic Interpretability Language Model Steering

Open paper AskPDF arXiv

Language models encode task-relevant knowledge in internal representations that far exceeds their output performance, but whether mechanistic interpretability methods can bridge this knowledge-action gap has not been systematically tested. We compared four mechanistic interpretability methods -- concept bottleneck steering (Steerling-8B), sparse autoencoder feature steering, logit lens with activation patching, and linear probing with truthfulness separator vector steering (Qwen 2.5 7B Instruct) -- for correcting false-negative triage errors using 400 physician-adjudicated clinical vignettes (144 hazards, 256 benign). Linear probes discriminated hazardous from benign cases with 98.2% AUROC, yet the model's output sensitivity was only 45.1%, a 53-percentage-point knowledge-action gap. Concept bottleneck steering corrected 20% of missed hazards but disrupted 53% of correct detections, indistinguishable from random perturbation (p=0.84). SAE feature steering produced zero effect despite 3,695 significant features. TSV steering at high strength corrected 24% of missed hazards while disrupting 6% of correct detections, but left 76% of errors uncorrected. Current mechanistic interpretability methods cannot reliably translate internal knowledge into corrected outputs, with implications for AI safety frameworks that assume interpretability enables effective error correction.