English
Related papers

Related papers: Steered LLM Activations are Non-Surjective

200 papers

Large language models (LLMs) have shown remarkable success in recent years, enabling a wide range of applications, including intelligent assistants that support users' daily life and work. A critical factor in building such assistants is…

Computation and Language · Computer Science 2025-10-28 Xiaoyan Zhao , Ming Yan , Yilun Qiu , Haoting Ni , Yang Zhang , Fuli Feng , Hong Cheng , Tat-Seng Chua

Interaction with Large Language Models (LLMs) is primarily carried out via prompting. A prompt is a natural language instruction designed to elicit certain behaviour or output from a model. In theory, natural language prompts enable…

Human-Computer Interaction · Computer Science 2024-03-15 Michael Desmond , Michelle Brachman

Steering language models (LMs) by modifying internal activations is a popular approach for controlling text generation. Unsupervised dictionary learning methods, e.g., sparse autoencoders, can be scaled to produce many steering vectors, but…

Computation and Language · Computer Science 2025-06-05 Jiuding Sun , Sidharth Baskaran , Zhengxuan Wu , Michael Sklar , Christopher Potts , Atticus Geiger

Fine-tuning large language models (LLMs) to adapt to evolving safety policies is costly and impractical. Mechanistic interpretability enables inference-time control through latent activation steering, yet its potential for precise,…

Machine Learning · Computer Science 2025-06-06 Shaona Ghosh , Amrita Bhattacharjee , Yftah Ziser , Christopher Parisien

Complex social behaviors, such as empathy and strategic politeness, are widely assumed to resist the directional decomposition that makes activation steering effective for coarse attributes like sentiment or toxicity. We present STAR:…

Computation and Language · Computer Science 2026-03-18 Niranjan Chebrolu , Kokil Jaidka , Gerard Christopher Yeo

Steering methods influence Large Language Model behavior by identifying semantic directions in hidden representations, but are typically realized through inference-time activation interventions that apply a fixed, global modification to the…

Computation and Language · Computer Science 2026-03-04 Chung-En Sun , Ge Yan , Zimo Wang , Tsui-Wei Weng

Activation steering has emerged as a powerful tool to shape LLM behavior without the need for weight updates. While its inherent brittleness and unreliability are well-documented, its safety implications remain underexplored. In this work,…

Cryptography and Security · Computer Science 2026-03-26 Yuxiao Li , Alina Fastowski , Efstratios Zaradoukas , Bardh Prenkaj , Gjergji Kasneci

Recent advances in mechanistic interpretability have revealed that large language models (LLMs) develop internal representations corresponding not only to concrete entities but also distinct, human-understandable abstract concepts and…

Machine Learning · Computer Science 2025-12-01 Rio Alexa Fear , Payel Mukhopadhyay , Michael McCabe , Alberto Bietti , Miles Cranmer

Large language models are highly sensitive to prompts, but this sensitivity is usually studied through task-relevant instructions, demonstrations, or reasoning cues. In this paper, we study a different form of prompt sensitivity: whether…

Computation and Language · Computer Science 2026-05-29 Pawel Batorski , Abtin Pourhadi , Jerzy Sarosiek , Przemyslaw Spurek , Paul Swoboda

A key challenge in AI alignment is guiding large language models (LLMs) to follow desired behaviors at test time. Activation steering, which modifies internal model activations during inference, offers a potential solution. However, prior…

Machine Learning · Computer Science 2025-03-04 Reza Bayat , Ali Rahimi-Kalahroudi , Mohammad Pezeshki , Sarath Chandar , Pascal Vincent

Algorithmic audits are essential tools for examining systems for properties required by regulators or desired by operators. Current audits of large language models (LLMs) primarily rely on black-box evaluations that assess model behavior…

Computers and Society · Computer Science 2026-05-19 Hannah Cyberey , Yangfeng Ji , David Evans

Changing the behavior of large language models (LLMs) can be as straightforward as editing the Transformer's residual streams using appropriately constructed "steering vectors." These modifications to internal neural activations, a form of…

Computation and Language · Computer Science 2025-05-20 Jian-Qiao Zhu , Haijiang Yan , Thomas L. Griffiths

Recent work has demonstrated the potential of contrastive steering for jailbreaking Large Language Models (LLMs). However, existing methods rely on limited and inherently biased contrastive prompts and require laborious manual tuning of…

Cryptography and Security · Computer Science 2026-05-21 Junxi Chen , Junhao Dong , Xiaohua Xie

Despite extensive efforts in safety alignment, large language models (LLMs) remain vulnerable to jailbreak attacks. Activation steering offers a training-free defense method but relies on fixed steering coefficients, resulting in suboptimal…

Cryptography and Security · Computer Science 2025-09-22 Weixiang Zhao , Jiahe Guo , Yulin Hu , Yang Deng , An Zhang , Xingyu Sui , Xinyang Han , Yanyan Zhao , Bing Qin , Tat-Seng Chua , Ting Liu

Post-training adaptation of language models is commonly achieved through parameter updates or input-based methods such as fine-tuning, parameter-efficient adaptation, and prompting. In parallel, a growing body of work modifies internal…

Computation and Language · Computer Science 2026-04-16 Simon Ostermann , Daniil Gurgurov , Tanja Baeumel , Michael A. Hedderich , Sebastian Lapuschkin , Wojciech Samek , Vera Schmitt

Reflection, the ability of large language models (LLMs) to evaluate and revise their own reasoning, has been widely used to improve performance on complex reasoning tasks. Yet, most prior works emphasizes designing reflective prompting…

Machine Learning · Computer Science 2025-12-12 Fu-Chieh Chang , Yu-Ting Lee , Pei-Yuan Wu

Activation steering methods enable inference-time control of large language model (LLM) behavior without retraining, but current approaches face a fundamental trade-off: sample-efficient methods suboptimally capture steering signals from…

Machine Learning · Computer Science 2026-03-09 Kartik Sharma , Rakshit S. Trivedi

Inference-time steering offers a promising way to control language models (LMs) without retraining. However, standard approaches typically rely on activation addition, which inevitably alters the hidden-state magnitudes raising concerns…

Machine Learning · Computer Science 2026-05-19 Zejia You , Chunyuan Deng , Hanjie Chen

Recent work in activation steering has demonstrated the potential to better control the outputs of Large Language Models (LLMs), but it involves finding steering vectors. This is difficult because engineers do not typically know how…

Computation and Language · Computer Science 2023-12-08 Ole Jorgensen , Dylan Cope , Nandi Schoots , Murray Shanahan

Controlling the behavior of large language models (LLMs) at inference time is essential for aligning outputs with human abilities and safety requirements. \emph{Activation steering} provides a lightweight alternative to prompt engineering…

Artificial Intelligence · Computer Science 2026-01-30 Diaoulé Diallo , Katharina Dworatzyk , Sophie Jentzsch , Peer Schütt , Sabine Theis , Tobias Hecking