Methods for controlling large language models (LLMs), including local weight fine-tuning, LoRA-based adaptation, and activation-based interventions, are often studied in isolation, obscuring their connections and making comparison difficult. In this work, we present a unified view that frames these interventions as dynamic weight updates induced by a control signal, placing them within a single conceptual framework. Building on this view, we propose a unified preference-utility analysis that separates control effects into preference, defined as the tendency toward a target concept, and utility, defined as coherent and task-valid generation, and measures both on a shared log-odds scale using polarity-paired contrastive examples. Across methods, we observe a consistent trade-off between preference and utility: stronger control increases preference while predictably reducing utility. We further explain this behavior through an activation manifold perspective, in which control shifts representations along target-concept directions to enhance preference, while utility declines primarily when interventions push representations off the model's valid-generation manifold. Finally, we introduce a new steering approach SPLIT guided by this analysis that improves preference while better preserving utility. Code is available at https://github.com/zjunlp/EasyEdit/blob/main/examples/SPLIT.md.
Multilingual Political Views of Large Language Models: Identification and Steering
Saarland University(萨尔兰州大学);University of Trento(特伦特大学);Brno University of Technology(布拉格技术大学);Kempelen Institute of Intelligent Technologies(凯姆佩尔智能技术研究所);German Research Center for AI (DFKI)(人工智能德国研究中心(DFKI))
Uncovering Why LLMs Fail: From Internal Mechanisms to Adversarial Reasoning Trajectories
Institute of Computing Technology, Chinese Academy of Sciences(中国科学院计算技术研究所);University of Chinese Academy of Sciences(中国科学院大学);State Key Laboratory of AI Safety, Institute of Computing Technology, Chinese Academy of Sciences(中国科学院计算技术研究所人工智能安全国家重点实验室);ShanghaiTech University(上海科技大学);Beihang University(北京航空航天大学)
ODESteer: A Unified ODE-Based Steering Framework for LLM Alignment
University of Illinois Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校);Northwestern University(西北大学);William & Mary(威廉与玛丽学院);University of Washington(华盛顿大学);Stanford University(斯坦福大学)
PIXEL: Adaptive Steering Via Position-wise Injection with eXact Estimated Levels under Subspace Calibration
University of Queensland(昆士兰大学);Mohamed bin Zayed University of Artificial Intelligence (MBZUAI)(穆罕默德·本·扎耶德人工智能大学);Provable Responsible AI and Data Analytics (PRADA) Lab(可证责任AI与数据分析师实验室);King Abdullah University of Science and Technology (KAUST)(国王 Abdullah 科学与技术大学)
A Hopfieldian View-based Interpretation for Chain-of-Thought Reasoning
KAUST(阿卜杜拉国王科技大学);Soochow University(苏州大学);University of Southampton(南安普顿大学);Tongji University(同济大学);Universität Hamburg(汉堡大学);Huazhong University of Science and Technology(华中科技大学)
Where Steering Signals Come From: Activation Source Selection in Activation Steering
Peking University(北京大学);Tsinghua University(清华大学);Beijing Academy of Artificial Intelligence(北京智源人工智能研究院);Tsinghua Shenzhen International Graduate School(清华大学深圳国际研究生院);YiXin-AILab(易信人工智能实验室)
Role Steering of Language Models for Social Simulations
Georgia Institute of Technology(佐治亚理工学院);ML Alignment & Theory Scholars (MATS)(ML对齐与理论学者组织(MATS));University of Illinois Urbana-Champaign(伊利诺伊大学厄巴纳-香槟分校);University of Oxford(牛津大学);Google DeepMind(谷歌DeepMind);OpenAI(开放人工智能公司)