DK

David Krueger

0000-0001-7256-0937

cs.LGcs.AIcs.CYstat.MLcs.CLcs.CVcs.CRcs.NEcond-mat.stat-mechcs.IT

On Valency

published · living versions

Monitorability Is Trainable: Post-Training for Chain-of-Thought Monitorability

with Usman Anwar, Tim Bakker, Christos Louizos

1 version

W_938rdgm8v1 · current

Preprints & journals

85 papers in the corpus · 2015–2026

Monitorability Is Trainable: Post-Training for Chain-of-Thought Monitorability

Usman Anwar, Tim Bakker, Christos Louizos et al.

0 citations

2610.29734v12026arXiv

Training LLMs to Verbalize Evaluation Awareness

Usman Anwar, Sahar Abdelnabi, David Krueger

0 citations

2609.36316v12026arXiv

Unified Neural Scaling Laws

Ethan Caballero, Priyank Jaini, David Krueger et al.

0 citations

2605.26248v12026arXiv

A Decision-Theoretic Formalisation of Steganography With Applications to LLM Monitoring

Usman Anwar, Julianna Piskorz, David D. Baek et al.

0 citations

2602.23163v32026arXiv

AI Researchers' Views on Automating AI R&D and Intelligence Explosions

Severin Field, Raymond Douglas, David Krueger

0 citations

2603.03338v22026arXiv

Position: Capability Control Should be a Separate Goal From Alignment

Shoaib Ahmed Siddiqui, Eleni Triantafillou, David Krueger et al.

0 citations

2602.05164v12026arXiv

Probabilistic Modelling is Sufficient for Causal Inference

Bruno Mlodozeniec, David Krueger, Richard E. Turner

0 citations

2512.23408v12025Proceedings of the 42nd International Conference on Machine Learning, PMLR 267:81810-81840, 2025

Fresh in memory: Training-order recency is linearly encoded in language model activations

Dmitrii Krasheninnikov, Richard E. Turner, David Krueger

0 citations

2509.14223v22025arXiv

Rethinking Safety in LLM Fine-tuning: An Optimization Perspective

Minseon Kim, Jin Myung Kwak, Lama Alssum et al.

0 citations

2508.12531v12025arXiv

How Do LLMs Persuade? Linear Probes Can Uncover Persuasion Dynamics in Multi-Turn Conversations

Brandon Jaipersaud, David Krueger, Ekdeep Singh Lubana

0 citations

2508.05625v12025arXiv

Mitigating Goal Misgeneralization via Minimax Regret

Karim Abdel Sadek, Matthew Farrugia-Roberts, Usman Anwar et al.

0 citations

2507.03068v22025arXiv

Distributional Training Data Attribution: What do Influence Functions Sample?

Bruno Mlodozeniec, Isaac Reid, Sam Power et al.

0 citations

2506.12965v32025arXiv

Detecting High-Stakes Interactions with Activation Probes

Alex McKenzie, Urja Pawar, Phil Blandfort et al.

0 citations

2506.10805v42025arXiv

From Dormant to Deleted: Tamper-Resistant Unlearning Through Weight-Space Regularization

Shoaib Ahmed Siddiqui, Adrian Weller, David Krueger et al.

1 citation

2505.22310v22025arXiv

Understanding (Un)Reliability of Steering Vectors in Language Models

Joschka Braun, Carsten Eickhoff, David Krueger et al.

0 citations

2505.22637v12025arXiv

Interpreting Emergent Planning in Model-Free Reinforcement Learning

Thomas Bush, Stephen Chung, Usman Anwar et al.

0 citations

2504.01871v12025arXiv

Taxonomy, Opportunities, and Challenges of Representation Engineering for Large Language Models

Jan Wehner, Sahar Abdelnabi, Daniel Tan et al.

0 citations

2502.19649v52025arXiv

Pitfalls of Evidence-Based AI Policy

Stephen Casper, David Krueger, Dylan Hadfield-Menell

3 citations

2502.09618v62025arXiv

Gradual Disempowerment: Systemic Existential Risks from Incremental AI Development

Jan Kulveit, Raymond Douglas, Nora Ammann et al.

13 citations

2501.16946v22025arXiv

Open Problems in Machine Unlearning for AI Safety

Fazl Barez, Tingchen Fu, Ameya Prabhu et al.

0 citations

2501.04952v12025arXiv

Learning to Forget using Hypernetworks

Jose Miguel Lara Rangel, Stefan Schoepf, Jack Foster et al.

0 citations

2412.00761v12024arXiv

Comparing Bottom-Up and Top-Down Steering Approaches on In-Context Learning Tasks

Madeline Brumley, Joe Kwon, David Krueger et al.

0 citations

2411.07213v12024arXiv

Noisy Zero-Shot Coordination: Breaking The Common Knowledge Assumption In Zero-Shot Coordination Games

Usman Anwar, Ashish Pandian, Jia Wan et al.

0 citations

2411.04976v12024arXiv

Understanding In-Context Learning of Linear Models in Transformers Through an Adversarial Lens

Usman Anwar, Johannes Von Oswald, Louis Kirsch et al.

0 citations

2411.05189v22024arXiv

Enhancing Neural Network Interpretability with Feature-Aligned Sparse Autoencoders

Luke Marks, Alasdair Paren, David Krueger et al.

4 citations

2411.01220v22024arXiv

Predicting Future Actions of Reinforcement Learning Agents

Stephen Chung, Scott Niekum, David Krueger

3 citations

2410.22459v12024arXiv

Integrating uncertainty quantification into randomized smoothing based robustness guarantees

Sina Daubener, Kira Maag, David Krueger et al.

0 citations

2410.20432v12024arXiv

Towards Reliable Evaluation of Behavior Steering Interventions in LLMs

Itamar Pres, Laura Ruis, Ekdeep Singh Lubana et al.

0 citations

2410.17245v12024arXiv

Influence Functions for Scalable Data Attribution in Diffusion Models

Bruno Mlodozeniec, Runa Eschenhagen, Juhan Bae et al.

1 citation

2410.13850v52024arXiv

Analyzing (In)Abilities of SAEs via Formal Languages

Abhinav Menon, Manish Shrivastava, David Krueger et al.

0 citations

2410.11767v42024arXiv

PoisonBench: Assessing Large Language Model Vulnerability to Data Poisoning

Tingchen Fu, Mrinank Sharma, Philip Torr et al.

2 citations

2410.08811v22024arXiv

Quantifying Feature Space Universality Across Large Language Models via Sparse Autoencoders

Michael Lan, Philip Torr, Austin Meek et al.

2 citations

2410.06981v42024arXiv

Towards Interpreting Visual Information Processing in Vision-Language Models

Clement Neo, Luke Ong, Philip Torr et al.

3 citations

2410.07149v22024arXiv

Exploring the design space of deep-learning-based weather forecasting systems

Shoaib Ahmed Siddiqui, Jean Kossaifi, Boris Bonev et al.

3 citations

2410.07472v12024arXiv

Permissive Information-Flow Analysis for Large Language Models

Shoaib Ahmed Siddiqui, Radhika Gaonkar, Boris Kopf et al.

0 citations

2410.03055v32024arXiv

Input Space Mode Connectivity in Deep Neural Networks

Jakub Vrabel, Ori Shem-Ur, Yaron Oz et al.

0 citations

2409.05800v12024arXiv

Protecting against simultaneous data poisoning attacks

Neel Alex, Shoaib Ahmed Siddiqui, Amartya Sanyal et al.

1 citation

2408.13221v12024arXiv

A deeper look at depth pruning of LLMs

Shoaib Ahmed Siddiqui, Xin Dong, Greg Heinrich et al.

0 citations

2407.16286v12024arXiv

The Perils of Optimizing Learned Reward Functions: Low Training Error Does Not Guarantee Low Regret

Lukas Fluri, Leon Lang, Alessandro Abate et al.

1 citation

2406.15753v32024arXiv

IDs for AI Systems

Alan Chan, Noam Kolt, Peter Wills et al.

0 citations

2406.12137v32024arXiv

Stress-Testing Capability Elicitation With Password-Locked Models

Ryan Greenblatt, Fabien Roger, Dmitrii Krasheninnikov et al.

0 citations

2405.19550v12024arXiv

Managing extreme AI risks amid rapid progress.

Bengio, Yoshua, Hinton, Geoffrey, Yao, Andrew et al.

347 citations

387682792024Science (New York, N.Y.). 2024;384(6698):842-845

Foundational Challenges in Assuring Alignment and Safety of Large Language Models

Usman Anwar, Abulhair Saparov, Javier Rando et al.

19 citations

2404.09932v22024arXiv

Affirmative safety: An approach to risk management for high-risk AI

Akash R. Wasil, Joshua Clymer, David Krueger et al.

1 citation

2406.15371v12024arXiv

Safety Cases: How to Justify the Safety of Advanced AI Systems

Joshua Clymer, Nick Gabrieli, David Krueger et al.

5 citations

2403.10462v22024arXiv

A Generative Model of Symmetry Transformations

James Urquhart Allingham, Bruno Kacper Mlodozeniec, Shreyas Padhy et al.

7 citations

2403.01946v32024arXiv

Black-Box Access is Insufficient for Rigorous AI Audits

Stephen Casper, Carson Ezell, Charlotte Siegmann et al.

79 citations

2401.14446v32024The 2024 ACM Conference on Fairness, Accountability, and Transparency (FAccT '24), June 3-6, 2024, Rio de Janeiro, Brazil

Visibility into AI Agents

Alan Chan, Carson Ezell, Max Kaufmann et al.

81 citations

2401.13138v62024arXiv

Hazards from Increasingly Accessible Fine-Tuning of Downloadable Foundation Models

Alan Chan, Ben Bucknall, Herbie Bradley et al.

1 citation

2312.14751v12023arXiv

Mechanistically analyzing the effects of fine-tuning on procedurally defined tasks

Samyak Jain, Robert Kirk, Ekdeep Singh Lubana et al.

4 citations

2311.12786v22023arXiv

Career total: 100 works. 85 are in this corpus.Showing the 50 most recent.

Profile built from the corpus for this byline.

100

works

5,455

citations

33

h-index

Corpus papers per year

Author records are still filling in while the Hub is in beta. If this is your page, you'll be able to claim it soon. Spot a mistake? Tell us.