Overview

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

CardsList

417 papers59 papers in the last four weeks, down 9% on the four weeks before. 0.4% of all new papers.

  1. MAPLE: A Multi-Axis Parametric Knowledge Extraction Lens for LLM Evaluation

    Oct 14, 2026Minsuh Joo, Youna Kim, Choonghyun Park +2LLM EvaluationLLM Information Extraction

  2. Genesis Steering: Better Activation Steering Through Training Data

    Oct 13, 2026Tora Bodin, Eric Aubinais, Ilaria Torre +1Language Model SteeringActivation Steering

  3. Mitigating Cross-Modal Hallucinations by Correcting Modality Misrouting in Omnimodal LLMs

    Oct 13, 2026Yu Mi, Licheng Pan, Jiayi Zhu +7Multimodal Large Language ModelsMultimodal Hallucination

  4. Cross-Lingual Transfer of Evil Vectors Is Directionally Asymmetric

    Oct 13, 2026Juann Kim, Minku KangCross-Lingual TransferActivation Steering

  5. Beyond Bias Scores: Answerability Confounds in Activation Steering with Gradient-Constrained Direction Design (GCD)

    Oct 13, 2026Weitong Li, Donglin Song, Junkai Lin +1Activation SteeringLanguage Model Bias Evaluation

  6. When a Conceptor Complement Becomes Gain: Operator Geometry and Behavioral Attribution in Language Models

    Oct 12, 2026Shuhul Razdan, R BalamuruganLLM InterpretabilityActivation Steering

  7. REVEAL: Reverse Engineering Agentic Latents from Frontier-Model Traces

    Oct 12, 2026Yuetian Du, Meng Wang, Yucheng Wang +9Activation Steering

  8. The Sharp Directions Are Against You: Curvature Analysis of Activation Steering

    Oct 12, 2026Seyedarmin Azizi, Arya Fayyazi, Parsa Razmara +3Language Model SteeringActivation Steering

  9. Harmonic Representations in Symbolic Music Transformers are Localized and Controllable

    Oct 11, 2026Adi Silberschein, Megan Wei, Tamar Rott ShahamTransformer InterpretabilityControllable Music Generation

  10. RIVO: Lightweight Adaptive Rerouting for Deep Research Agents

    Oct 11, 2026Mingchen Li, Jiatan Huang, Hong YuDeep Research AgentsActivation Steering

  11. Mind Your Own Role: Reducing Role Overstepping for Effective and Efficient Multi-Agent Collaboration via Dynamic Role Activation Steering

    Oct 11, 2026Xingke Yang, Sicong Li, Lingfeng Yao +5Multi-Agent LLM SystemsMulti-Agent Collaboration

  12. TRACE: Temporal Risk Assessment and Concept Editing for Safer Language Generation

    Oct 11, 2026Ankur Garg, Xuemin Yu, Hassan Sajjad +1LLM SafetyLLM Safety Monitoring

  13. FleetPilot: Fine-Grained Steering Empowers Real-World LLM Personalization

    Oct 10, 2026Kounianhua Du, Jianxing Liu, Kangning Zhang +7Personalized Language ModelsLanguage Model Steering

  14. Steerspeech: Activation Steering For Emotion Control In Generated Speech

    Oct 7, 2026Afsara Benazir, Darius Pétermann, Felix Xiaozhu Lin +1Controllable Speech GenerationEmotional Speech Synthesis

  15. Nullify: Null-Space Activation Steering for Training-Free LLM Unlearning

    Oct 7, 2026Wei Zhai, Xiang Liu, Qiang Huang +6LLM UnlearningPrivacy Leakage in Language Models

  16. PatchBench: Measuring Collateral Damage in Activation Patching

    Oct 7, 2026Alexi Canesse, Mathis Le Bail, Maël Jenny +3Jailbreak Attack EvaluationLanguage Model Safety Evaluation

  17. The Attribution Blind Spot: Layerwise Trajectory Diagnostics for Source Reliance in Retrieval-Augmented Language Models

    Oct 7, 2026Zhe Yu, Wenpeng Xing, Yunzhao Wei +4LLM InterpretabilityRetrieval-Augmented Generation

  18. Does Steering Break Your Model? A Multi-Dimensional Evaluation Suite for LLM Steering Methods

    Oct 6, 2026Haotian Yang, Huikang Jiang, Yucheng Wu +4LLM EvaluationLanguage Model Steering

  19. Loud and Clear: Dynamic Activation Steering for Improving Speech Intelligibility in Noisy Environments

    Oct 6, 2026Seymanur Akti, Alexander WaibelTTS SynthesisControllable Speech Generation

  20. COMPASS: Finding Where Reasoning Lives in Language Models

    Oct 5, 2026Pratyay Dutta, Kowshik Thopalli, Vivek NarayanaswamyEfficient Language Model ReasoningActivation Steering

  21. Steering by Influence: Curvature Aware Data Weighting for Activation Steering

    Oct 5, 2026James A. E. Dixon, Stephen J. Roberts, Francesco QuinzanLanguage Model SteeringInfluence Functions

  22. TrustMI: Causally controlling how assistants trust their users

    Oct 5, 2026Théo Lasnier, Romain Froger, Maxence Lasbordes +1LLM SafetyTrust in AI

  23. Noise Out, Bias In: Targeted Bias Injection in Diffusion Language Models via Closed-Loop Activation Steering

    Oct 5, 2026Sarim Hashmi, Mukul Ranjan, Abdelrahman Elsayed +3Adversarial Attacks on Diffusion ModelsDiffusion Language Models

  24. Usage-Modulated Sentiment Representations in Large Language Models

    Oct 4, 2026Hongfei Du, Jiacheng Shi, Yanfu Zhang +2Causal Interventions in Language ModelsRepresentation Geometry in Language Models

  25. EmoRSS: Mitigating Emotion-Induced Over-Refusal in Large Language Models

    Oct 4, 2026Shuyi Miao, Yaojin Ma, Chenhang Cui +5LLM AlignmentLLM Refusal Behavior