Steering at the Source: Style Modulation Heads for Robust Persona Control
The University of Tokyo, Tokyo, Japan(东京大学);National Institute of Informatics Research(信息处理研究所);Development Center for Large Language Models, Japan(大型语言模型发展中心)
Activation steering offers a computationally efficient mechanism for controlling Large Language Models (LLMs) without fine-tuning. While effectively controlling target traits (e.g., persona), coherency degradation remains a major obstacle to safety and practical deployment. We hypothesize that this degradation stems from intervening on the residual stream, which indiscriminately affects aggregated features and inadvertently amplifies off-target noise. In this work, we identify a sparse subset of attention heads (only three heads) that independently govern persona and style formation, which we term Style Modulation Heads. Specifically, these heads can be localized via geometric analysis of internal representations, combining layer-wise cosine similarity and head-wise contribution scores. We demonstrate that intervention targeting only these specific heads achieves robust behavioral control while significantly mitigating the coherency degradation observed in residual stream steering. More broadly, our findings show that precise, component-level localization enables safer and more precise model control.
Do Personality Traits Interfere? Geometric Limitations of Steering in Large Language Models
Network Analysis and Social Influence Modelling (NASIM) Lab(网络分析与社会影响建模实验室);School of Physics Maths and Computing(物理数学计算学院);The University of Western Australia(西澳大学);School of Computing(计算学院);Macquarie University(麦觉瑞大学)
A Systematic Analysis of the Impact of Persona Steering on LLM Capabilities
School of Computer Science and Engineering, Northeastern University, Shenyang 110819, China(东北大学计算机科学与工程学院,沈阳 110819,中国);School of Computing and Information Systems, Singapore Management University, Singapore 178902, Singapore(新加坡管理大学计算机与信息系统学院,新加坡 178902,新加坡);School of Computer and Communication Engineering, Northeastern University, Qinhuangdao 066004, China(东北大学计算机与通信工程学院,秦皇岛 066004,中国)
BILLY: Steering Large Language Models via Merging Persona Vectors for Creative Generation
Department of Electrical Engineering, National Taiwan University(国立台湾大学电子工程系);Department of Computer Science & Information Engineering, National Taiwan University(国立台湾大学计算机科学与信息工程系);Graduate Institute of Communication Engineering, National Taiwan University(国立台湾大学通信工程研究所);CSAIL, Massachusetts Institute of Technology(麻省理工学院计算机科学与人工智能实验室)
Activation-Space Personality Steering: Hybrid Layer Selection for Stable Trait Control in LLMs
Network Analysis and Social Influence Modelling (NASIM) Lab(网络分析与社会影响建模实验室);School of Physics Maths and Computing, The University of Western Australia(西澳大学物理数学与计算学院);School of Psychological Science, The University of Western Australia(西澳大学心理学科学学院);School of Computing, Macquarie University(麦考瑞大学计算机学院)
Training-Free Cultural Alignment of Large Language Models via Persona Disagreement
Faculty of Information and Technology, University of Science, Vietnam National University(信息与技术学院,科学大学,越南国家大学);Department of Computer Science, University of Warwick(计算机科学系,沃里克大学);School of Computing, Engineering and Digital Technologies, Teesside University(计算、工程和数字技术学院,泰赛德大学)
LF-Steering: Latent Feature Activation Steering for Enhancing Semantic Consistency in Large Language Models
College of Intelligence and Computing, Tianjin University(天津大学智能与计算学部);IT Innovation and Research Center, Huawei Technologies(华为技术有限公司 IT 创新研究中心);Informatics, University of Edinburgh(爱丁堡大学信息学院)