English
Related papers

Related papers: Activation Scaling for Steering and Interpreting L…

200 papers

Research in machine learning has successfully developed algorithms to build accurate classification models. However, in many real-world applications, such as healthcare, customer satisfaction, and environment protection, we want to be able…

Machine Learning · Computer Science 2020-12-08 Samuel Marc Denton , Ansaf Salleb-Aouissi

Transformers achieve unrivalled performance in modelling language, but remain inefficient in terms of memory and time complexity. A possible remedy is to reduce the sequence length in the intermediate layers by pooling fixed-length segments…

Computation and Language · Computer Science 2023-10-25 Piotr Nawrot , Jan Chorowski , Adrian Łańcucki , Edoardo M. Ponti

Large language models (LLMs) exhibit persistent miscalibration, especially after instruction tuning and preference alignment. Modified training objectives can improve calibration, but retraining is expensive. Inference-time steering offers…

Machine Learning · Computer Science 2026-02-06 Miranda Muqing Miao , Young-Min Cho , Lyle Ungar

Safety alignment is indispensable for Large Language Models (LLMs) to defend threats from malicious instructions. However, recent researches reveal safety-aligned LLMs prone to reject benign queries due to the exaggerated safety issue,…

Artificial Intelligence · Computer Science 2024-12-18 Zouying Cao , Yifei Yang , Hai Zhao

The mechanisms by which reasoning training reshapes LLMs' internal computations remain unclear. We study lightweight steering vectors inserted into the base model's residual stream and trained with a reinforcement-learning objective. These…

There has been increasing interest in evaluations of language models for a variety of risks and characteristics. Evaluations relying on natural language understanding for grading can often be performed at scale by using other language…

Computation and Language · Computer Science 2023-12-11 Simon Lermen , Ondřej Kvapil

Despite significant progress in alignment, large language models (LLMs) remain vulnerable to adversarial attacks that elicit harmful behaviors. Activation steering techniques offer a promising inference-time intervention approach, but…

Machine Learning · Computer Science 2026-01-28 Quy-Anh Dang , Chris Ngo

Building pluralistic AI requires designing models that are able to be shaped to represent a wide range of value systems and cultures. Achieving this requires first being able to evaluate the degree to which a given model is capable of…

Teacher education requires deliberate practice with learners who exhibit identifiable strengths, weaknesses, and partial mastery. Large language models could support such practice by simulating students with known skill components, enabling…

Computation and Language · Computer Science 2026-05-26 Alexander Apartsin , Omri Sason , Yehudit Aperstein

Accounts of human language processing have long appealed to implicit ``situation models'' that enrich comprehension with relevant but unstated world knowledge. Here, we apply causal intervention techniques to recent transformer models to…

Computation and Language · Computer Science 2023-06-08 Takateru Yamakoshi , James L. McClelland , Adele E. Goldberg , Robert D. Hawkins

Some actions must be executed in different ways depending on the context. For example, wiping away marker requires vigorous force while wiping away almonds requires more gentle force. In this paper we provide a model where an agent learns…

Computation and Language · Computer Science 2023-02-08 Mattias Appelgren , Alex Lascarides

The increasing adoption of machine learning tools has led to calls for accountability via model interpretability. But what does it mean for a machine learning model to be interpretable by humans, and how can this be assessed? We focus on…

Machine Learning · Computer Science 2019-08-06 Dylan Slack , Sorelle A. Friedler , Carlos Scheidegger , Chitradeep Dutta Roy

Large language models (LLMs) can be controlled at inference time through prompts (in-context learning) and internal activations (activation steering). Different accounts have been proposed to explain these methods, yet their common goal of…

Machine Learning · Computer Science 2026-03-13 Eric Bigelow , Daniel Wurgaft , YingQiao Wang , Noah Goodman , Tomer Ullman , Hidenori Tanaka , Ekdeep Singh Lubana

Large language models possess strong chemical reasoning capabilities, making them effective molecular editors. However, property-relevant information is implicitly entangled across their dense hidden states, providing no explicit handle for…

Machine Learning · Computer Science 2026-05-12 Mingxu Zhang , Yuhan Li , Lujundong Li , Dazhong Shen , Hui Xiong , Ying Sun

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

Different activation functions work best for different deep learning models. To exploit this, we leverage recent advancements in gradient-based search techniques for neural architectures to efficiently identify high-performing activation…

Machine Learning · Computer Science 2024-08-14 Lukas Strack , Mahmoud Safari , Frank Hutter

With the advancement in generative language models, the selection of prompts has gained significant attention in recent years. A prompt is an instruction or description provided by the user, serving as a guide for the generative language…

Machine Learning · Statistics 2024-05-21 Haoting Zhang , Jinghai He , Rhonda Righter , Zeyu Zheng

Sampling-based search, a simple paradigm for utilizing test-time compute, involves generating multiple candidate responses and selecting the best one -- typically by having models self-verify each response for correctness. In this paper, we…

Machine Learning · Computer Science 2025-02-21 Eric Zhao , Pranjal Awasthi , Sreenivas Gollapudi

An unintended consequence of the vast pretraining of Large Language Models (LLMs) is the verbatim memorization of fragments of their training data, which may contain sensitive or copyrighted information. In recent years, unlearning has…

Computation and Language · Computer Science 2024-11-06 Atakan Seyitoğlu , Aleksei Kuvshinov , Leo Schwinn , Stephan Günnemann

The performance of active learning algorithms can be improved in two ways. The often used and intuitive way is by reducing the overall error rate within the test set. The second way is to ensure that correct predictions are not forgotten…

Machine Learning · Computer Science 2024-11-19 Ryan Benkert , Mohit Prabhushankar , Ghassan AlRegib