English
Related papers

Related papers: Steered LLM Activations are Non-Surjective

200 papers

Chain-of-Thought (CoT) prompting has improved the reasoning performance of large language models (LLMs), but it remains unclear why it works and whether it is the unique mechanism for triggering reasoning in large language models. In this…

Computation and Language · Computer Science 2026-01-14 Zhenghao He , Guangzhi Xiong , Bohan Liu , Sanchit Sinha , Aidong Zhang

Language models (LMs) are increasingly used in high-stakes, multi-agent settings, where following instructions and maintaining value alignment are critical. Most alignment research focuses on interactions between a single LM and a single…

Artificial Intelligence · Computer Science 2026-05-12 Maria Chang , Ronny Luss , Miao Liu , Keerthiram Murugesan , Karthikeyan Ramamurthy , Djallel Bouneffouf

Protein Language Models (PLMs), pre-trained on extensive evolutionary data from natural proteins, have emerged as indispensable tools for protein design. While powerful, PLMs often struggle to produce proteins with precisely specified…

Biomolecules · Quantitative Biology 2025-09-15 Long-Kai Huang , Rongyi Zhu , Bing He , Jianhua Yao

Activation steering controls language model behavior by adding directions to internal representations at inference time, but standard residual-stream steering can fail in stateful dialogue. We identify KV-cache contamination as a key…

Computation and Language · Computer Science 2026-05-15 Diancheng Kang , Zheyuan Liu , Ningshan Ma , Yue Huang , Zhaoxuan Tan , Meng Jiang

Safety-aligned large language models (LLMs) are increasingly deployed in real-world pipelines, yet this deployment also enlarges the supply-chain attack surface: adversaries can distribute backdoored checkpoints that behave normally under…

Cryptography and Security · Computer Science 2026-04-15 Rui Yin , Tianxu Han , Naen Xu , Changjiang Li , Ping He , Chunyi Zhou , Jun Wang , Zhihui Fu , Tianyu Du , Jinbao Li , Shouling Ji

Large language models (LLMs) can sometimes detect when they are being evaluated and adjust their behavior to appear more aligned, compromising the reliability of safety evaluations. In this paper, we show that adding a steering vector to an…

Computation and Language · Computer Science 2026-03-03 Tim Tian Hua , Andrew Qin , Samuel Marks , Neel Nanda

Large language models (LLMs) have shown impressive abilities in leveraging pretrained knowledge through prompting, but they often struggle with unseen tasks, particularly in data-scarce scenarios. While cross-task in-context learning offers…

Computation and Language · Computer Science 2025-07-18 Xinyu Tang , Zhihao Lv , Xiaoxue Cheng , Junyi Li , Wayne Xin Zhao , Zujie Wen , Zhiqiang Zhang , Jun Zhou

Language models (LMs) are typically post-trained for desired capabilities and behaviors via weight-based or prompt-based steering, but the former is time-consuming and expensive, and the latter is not precisely controllable and often…

Computation and Language · Computer Science 2026-05-18 Sasha Cui , Zhongren Chen

Modern large language models (LLMs) are typically secured by auditing data, prompts, and refusal policies, while treating the forward pass as an implementation detail. We show that intermediate activations in decoder-only LLMs form a…

Cryptography and Security · Computer Science 2025-11-24 Zhiyuan Xu , Stanislav Abaimov , Joseph Gardiner , Sana Belguith

This work introduces SteerVLM, a lightweight steering module designed to guide Vision-Language Models (VLMs) towards outputs that better adhere to desired instructions. Our approach learns from the latent embeddings of paired prompts…

Computer Vision and Pattern Recognition · Computer Science 2025-10-31 Anushka Sivakumar , Andrew Zhang , Zaber Hakim , Chris Thomas

As LLMs are increasingly deployed in real-world applications, ensuring their ability to refuse malicious prompts, especially jailbreak attacks, is essential for safe and reliable use. Recently, activation steering has emerged as an…

Machine Learning · Computer Science 2026-02-10 Leheng Sheng , Changshuo Shen , Weixiang Zhao , Junfeng Fang , Xiaohao Liu , Zhenkai Liang , Xiang Wang , An Zhang , Tat-Seng Chua

We introduce Contrastive Activation Addition (CAA), an innovative method for steering language models by modifying their activations during forward passes. CAA computes "steering vectors" by averaging the difference in residual stream…

Computation and Language · Computer Science 2024-07-08 Nina Panickssery , Nick Gabrieli , Julian Schulz , Meg Tong , Evan Hubinger , Alexander Matt Turner

Activation steering has emerged as a promising alternative for controlling language-model behavior at inference time by modifying intermediate representations while keeping model parameters frozen. However, large-scale evaluations such as…

Computation and Language · Computer Science 2026-05-08 Zehao Jin , Ruixuan Deng , Junran Wang , Xinjie Shen , Chao Zhang

LLMs are deployed globally, yet produce responses biased towards cultures with abundant training data. Existing cultural localization approaches such as prompting or post-training alignment are black-box, hard to control, and do not reveal…

Computation and Language · Computer Science 2026-03-25 Simran Khanuja , Hongbin Liu , Shujian Zhang , John Lambert , Mingqing Chen , Rajiv Mathews , Lun Wang

Adapting models to a language that was only partially present in the pre-training data requires fine-tuning, which is expensive in terms of both data and computational resources. As an alternative to fine-tuning, we explore the potential of…

Computation and Language · Computer Science 2024-11-28 Daniel Scalena , Elisabetta Fersini , Malvina Nissim

Recent advances in large language models (LLMs) have led to the development of thinking language models that generate extensive internal reasoning chains before producing responses. While these models achieve improved performance,…

Machine Learning · Computer Science 2025-10-23 Constantin Venhoff , Iván Arcuschin , Philip Torr , Arthur Conmy , Neel Nanda

Reliable behavior control is central to deploying large language models (LLMs) on the web. Activation steering offers a tuning-free route to align attributes (e.g., truthfulness) that ensure trustworthy generation. Prevailing approaches…

Artificial Intelligence · Computer Science 2025-11-19 Manjiang Yu , Hongji Li , Priyanka Singh , Xue Li , Di Wang , Lijie Hu

Steering, or direct manipulation of internal activations to guide LLM responses toward specific semantic concepts, is emerging as a promising avenue for both understanding how semantic concepts are stored within LLMs and advancing LLM…

Machine Learning · Computer Science 2026-02-03 Parmida Davarmanesh , Ashia Wilson , Adityanarayanan Radhakrishnan

Large Language Models (LLMs) exhibit remarkable capabilities across various tasks, yet guiding them to follow desired behaviours during inference remains a significant challenge. Activation steering offers a promising method to control the…

Computation and Language · Computer Science 2025-09-29 Weixuan Wang , Minghao Wu , Barry Haddow , Alexandra Birch

Activation engineering is becoming increasingly popular as a means of online control of large language models (LLMs). In this work, we extend the idea of inference-time steering with vectors that represent a behavioral direction of interest…

Machine Learning · Computer Science 2024-11-26 Christopher M. Ackerman