Maksym Andriushchenko

71 papers A* 19A 3Journal 43Unranked 5

Year	Rank	Type	Title / Venue / Authors
2026	J	`jnl`	Claudini: Autoresearch Discovers State-of-the-Art Adversarial Attack Algorithms for LLMs. CoRR Alexander Panfilov, Peter Romov, Igor Shilov, Yves-Alexandre de Montjoye, Jonas Geiping, Maksym Andriushchenko
2026	J	`jnl`	HalluHard: A Hard Multi-Turn Hallucination Benchmark. CoRR Dongyang Fan, Sebastien Delsad, Nicolas Flammarion, Maksym Andriushchenko
2026	J	`jnl`	Helpful to a Fault: Measuring Illicit Assistance in Multi-Turn, Multilingual LLM Agents. CoRR Nivya Talokar, Ayush K. Tarun, Murari Mandal, Maksym Andriushchenko, Antoine Bosselut
2026	J	`jnl`	International AI Safety Report 2026. CoRR Yoshua Bengio, Stephen Clare, Carina Prunkl, Maksym Andriushchenko, Ben Bucknall, Malcolm Murray, Rishi Bommasani, Stephen Casper, Tom Davidson, Raymond Douglas, David Duvenaud, Philip Fox, Usman Gohar, Rose Hadshar, Anson Ho, Tiancheng Hu, Cameron Jones, Sayash Kapoor, Atoosa Kasirzadeh, Sam Manning, Nestor Maslej, Vasilios Mavroudis, Conor McGlynn, Richard Moulange, Jessica Newman, Kwan Yee Ng, Patricia Paskov, Shalaleh Rismani, Girish Sastry, Elizabeth Seger, Scott Singer, Charlotte Stix, Lucia Velasco, Nicole Wheeler
2026	J	`jnl`	PostTrainBench: Can LLM Agents Automate LLM Post-Training? CoRR Ben Rank, Hardik Bhatnagar, Ameya Prabhu, Shira Eisenberg, Karina Nguyen, Matthias Bethge, Maksym Andriushchenko
2026	J	`jnl`	Skill-Inject: Measuring Agent Vulnerability to Skill File Attacks. CoRR David Schmotz, Luca Beurer-Kellner, Sahar Abdelnabi, Maksym Andriushchenko
2025	J	`jnl`	Adaptive Attacks on Trusted Monitors Subvert AI Control Protocols. CoRR Mikhail Terekhov, Alexander Panfilov, Daniil Dzenhaliou, Caglar Gulcehre, Maksym Andriushchenko, Ameya Prabhu, Jonas Geiping
2025	J	`jnl`	Agent Skills Enable a New Class of Realistic and Trivially Simple Prompt Injections. CoRR David Schmotz, Sahar Abdelnabi, Maksym Andriushchenko
2025	A*	`conf`	AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents. ICLR Maksym Andriushchenko, Alexandra Souly, Mateusz Dziemian, Derek Duenas, Maxwell Lin, Justin Wang, Dan Hendrycks, Andy Zou, J. Zico Kolter, Matt Fredrikson, Yarin Gal, Xander Davies
2025	J	`jnl`	Capability-Based Scaling Laws for LLM Red-Teaming. CoRR Alexander Panfilov, Paul Kassianik, Maksym Andriushchenko, Jonas Geiping
2025	A	`conf`	Critical Influence of Overparameterization on Sharpness-aware Minimization. UAI Sungbin Shin, Dongyeop Lee, Maksym Andriushchenko, Namhoon Lee
2025	A*	`conf`	Does Refusal Training in LLMs Generalize to the Past Tense? ICLR Maksym Andriushchenko, Nicolas Flammarion
2025	J	`jnl`	International AI Safety Report 2025: First Key Update: Capabilities and Risk Implications. CoRR Yoshua Bengio, Stephen Clare, Carina Prunkl, Shalaleh Rismani, Maksym Andriushchenko, Ben Bucknall, Philip Fox, Tiancheng Hu, Cameron Jones, Sam Manning, Nestor Maslej, Vasilios Mavroudis, Conor McGlynn, Malcolm Murray, Charlotte Stix, Lucia Velasco, Nicole Wheeler, Daniel Privitera, Sören Mindermann, Daron Acemoglu, Thomas G. Dietterich, Fredrik Heintz, Geoffrey E. Hinton, Nicholas R. Jennings, Susan Leavy, Teresa Ludermir, Vidushi Marda, Helen Margetts, John A. McDermid, Jane Munga, Arvind Narayanan, Alondra Nelson, Clara Neppel, Gopal Ramchurn, Stuart Russell, Marietje Schaake, Bernhard Schölkopf, Alavaro Soto, Lee Tiedrich, Gaël Varoquaux, Andrew Yao, Yan-Ning Zhang, Leandro Aguirre, Olubunmi Ajala, Fahad Albalawi Noora AlMalek, Christian Busch, André C. P. L. F. de Carvalho, Jonathan Collas, Amandeep Gill, Ahmet Halit Hatip, Juha Heikkilä, Chris Johnson, Gill Jolly, Ziv Katzir, Mary Kerema, Hiroaki Kitano, Antonio Krüger, Aoife McLysaght, Oleksii Molchanovskyi, Andrea Monti, Kyoung Mu Lee, Mona Nemer, Nuria Oliver, Raquel Pezoa Rivera, Audrey Plonk, José Ramón López Portillo, Balaraman Ravindran, Hammam Riza, Crystal Rugege, Haroon Sheikh, Denise Wong, Yi Zeng, Liming Zhu
2025	J	`jnl`	International AI Safety Report 2025: Second Key Update: Technical Safeguards and Risk Management. CoRR Yoshua Bengio, Stephen Clare, Carina Prunkl, Maksym Andriushchenko, Ben Bucknall, Philip Fox, Nestor Maslej, Conor McGlynn, Malcolm Murray, Shalaleh Rismani, Stephen Casper, Jessica Newman, Daniel Privitera, Sören Mindermann, Daron Acemoglu, Thomas G. Dietterich, Fredrik Heintz, Geoffrey E. Hinton, Nick R. Jennings, Susan Leavy, Teresa Ludermir, Vidushi Marda, Helen Margetts, John A. McDermid, Jane Munga, Arvind Narayanan, Alondra Nelson, Clara Neppel, Gopal Ramchurn, Stuart Russell, Marietje Schaake, Bernhard Schölkopf, Alavaro Soto, Lee Tiedrich, Gaël Varoquaux, Andrew Yao, Yanning Zhang, Leandro Aguirre, Olubunmi Ajala, Fahad Albalawi Noora AlMalek, Noora AlMalek, Christian Busch, André Carvalho, Jonathan Collas, Amandeep Gill, Ahmet Halit Hatip, Juha Heikkilä, Chris Johnson, Gill Jolly, Ziv Katzir, Mary Kerema, Hiroaki Kitano, Antonio Krüger, Aoife McLysaght, Oleksii Molchanovskyi, Andrea Monti, Kyoung Mu Lee, Mona Nemer, Nuria Oliver, Raquel Pezoa, Audrey Plonk, José Ramón López Portillo, Balaraman Ravindran, Hammam Riza, Crystal Rugege, Haroon Sheikh, Denise Wong, Yi Zeng, Liming Zhu
2025	A*	`conf`	Is In-Context Learning Sufficient for Instruction Following in LLMs? ICLR Hao Zhao, Maksym Andriushchenko, Francesco Croce, Nicolas Flammarion
2025	A*	`conf`	Jailbreaking Leading Safety-Aligned LLMs with Simple Adaptive Attacks. ICLR Maksym Andriushchenko, Francesco Croce, Nicolas Flammarion
2025	J	`jnl`	Monitoring Decomposition Attacks in LLMs with Lightweight Sequential Monitors. CoRR Chen Yueh-Han, Nitish Joshi, Yulin Chen, Maksym Andriushchenko, Rico Angell, He He
2025	J	`jnl`	OS-Harm: A Benchmark for Measuring Safety of Computer Use Agents. CoRR Thomas Kuntz, Agatha Duzan, Hao Zhao, Francesco Croce, Zico Kolter, Nicolas Flammarion, Maksym Andriushchenko
2025	J	`jnl`	Strategic Dishonesty Can Undermine AI Safety Evaluations of Frontier LLMs. CoRR Alexander Panfilov, Evgenii Kortukov, Kristina Nikolic, Matthias Bethge, Sebastian Lapuschkin, Wojciech Samek, Ameya Prabhu, Maksym Andriushchenko, Jonas Geiping
2024	J	`jnl`	AgentHarm: A Benchmark for Measuring Harmfulness of LLM Agents. CoRR Maksym Andriushchenko, Alexandra Souly, Mateusz Dziemian, Derek Duenas, Maxwell Lin, Justin Wang, Dan Hendrycks, Andy Zou, Zico Kolter, Matt Fredrikson, Eric Winsor, Jerome Wynne, Yarin Gal, Xander Davies
2024	J	`jnl`	Competition Report: Finding Universal Jailbreak Backdoors in Aligned LLMs. CoRR Javier Rando, Francesco Croce, Krystof Mitka, Stepan Shabalin, Maksym Andriushchenko, Nicolas Flammarion, Florian Tramèr
2024	J	`jnl`	Does Refusal Training in LLMs Generalize to the Past Tense? CoRR Maksym Andriushchenko, Nicolas Flammarion
2024	J	`jnl`	Exploring Memorization and Copyright Violation in Frontier LLMs: A Study of the New York Times v. OpenAI 2023 Lawsuit. CoRR Joshua Freeman, Chloe Rippe, Edoardo Debenedetti, Maksym Andriushchenko
2024	A*	`conf`	Improving Alignment and Robustness with Circuit Breakers. NeurIPS Andy Zou, Long Phan, Justin Wang, Derek Duenas, Maxwell Lin, Maksym Andriushchenko, J. Zico Kolter, Matt Fredrikson, Dan Hendrycks
2024	J	`jnl`	Improving Alignment and Robustness with Circuit Breakers. CoRR Andy Zou, Long Phan, Justin Wang, Derek Duenas, Maxwell Lin, Maksym Andriushchenko, Rowan Wang, Zico Kolter, Matt Fredrikson, Dan Hendrycks
2024	J	`jnl`	Is In-Context Learning Sufficient for Instruction Following in LLMs? CoRR Hao Zhao, Maksym Andriushchenko, Francesco Croce, Nicolas Flammarion
2024	A*	`conf`	JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Models. NeurIPS Patrick Chao, Edoardo Debenedetti, Alexander Robey, Maksym Andriushchenko, Francesco Croce, Vikash Sehwag, Edgar Dobriban, Nicolas Flammarion, George J. Pappas, Florian Tramèr, Hamed Hassani, Eric Wong
2024	J	`jnl`	JailbreakBench: An Open Robustness Benchmark for Jailbreaking Large Language Models. CoRR Patrick Chao, Edoardo Debenedetti, Alexander Robey, Maksym Andriushchenko, Francesco Croce, Vikash Sehwag, Edgar Dobriban, Nicolas Flammarion, George J. Pappas, Florian Tramèr, Hamed Hassani, Eric Wong
2024	J	`jnl`	Jailbreaking Leading Safety-Aligned LLMs with Simple Adaptive Attacks. CoRR Maksym Andriushchenko, Francesco Croce, Nicolas Flammarion
2024	A*	`conf`	Layer-wise linear mode connectivity. ICLR Linara Adilova, Maksym Andriushchenko, Michael Kamp, Asja Fischer, Martin Jaggi
2024	A*	`conf`	Long Is More for Alignment: A Simple but Tough-to-Beat Baseline for Instruction Fine-Tuning. ICML Hao Zhao, Maksym Andriushchenko, Francesco Croce, Nicolas Flammarion
2024	J	`jnl`	Long Is More for Alignment: A Simple but Tough-to-Beat Baseline for Instruction Fine-Tuning. CoRR Hao Zhao, Maksym Andriushchenko, Francesco Croce, Nicolas Flammarion
2024	—	`—`	Understanding generalization and robustness in modern deep learning. Maksym Andriushchenko
2024	A*	`conf`	Why Do We Need Weight Decay in Modern Deep Learning? NeurIPS Francesco D'Angelo, Maksym Andriushchenko, Aditya Vardhan Varre, Nicolas Flammarion
2023	A*	`conf`	A Modern Look at the Relationship between Sharpness and Generalization. ICML Maksym Andriushchenko, Francesco Croce, Maximilian Müller, Matthias Hein, Nicolas Flammarion
2023	J	`jnl`	A modern look at the relationship between sharpness and generalization. CoRR Maksym Andriushchenko, Francesco Croce, Maximilian Müller, Matthias Hein, Nicolas Flammarion
2023	A*	`conf`	SGD with Large Step Sizes Learns Sparse Features. ICML Maksym Andriushchenko, Aditya Vardhan Varre, Loucas Pillaud-Vivien, Nicolas Flammarion
2023	J	`jnl`	Scaling Compute Is Not All You Need for Adversarial Robustness. CoRR Edoardo Debenedetti, Zishen Wan, Maksym Andriushchenko, Vikash Sehwag, Kshitij Bhardwaj, Bhavya Kailkhura
2023	A*	`conf`	Sharpness-Aware Minimization Leads to Low-Rank Features. NeurIPS Maksym Andriushchenko, Dara Bahri, Hossein Mobahi, Nicolas Flammarion
2023	J	`jnl`	Sharpness-Aware Minimization Leads to Low-Rank Features. CoRR Maksym Andriushchenko, Dara Bahri, Hossein Mobahi, Nicolas Flammarion
2023	J	`jnl`	The Effects of Overparameterization on Sharpness-aware Minimization: An Empirical and Theoretical Analysis. CoRR Sungbin Shin, Dongyeop Lee, Maksym Andriushchenko, Namhoon Lee
2023	A*	`conf`	Transferable Adversarial Robustness for Categorical Data via Universal Robust Embeddings. NeurIPS Klim Kireev, Maksym Andriushchenko, Carmela Troncoso, Nicolas Flammarion
2023	J	`jnl`	Transferable Adversarial Robustness for Categorical Data via Universal Robust Embeddings. CoRR Klim Kireev, Maksym Andriushchenko, Carmela Troncoso, Nicolas Flammarion
2023	J	`jnl`	Why Do We Need Weight Decay in Modern Deep Learning? CoRR Maksym Andriushchenko, Francesco D'Angelo, Aditya Varre, Nicolas Flammarion
2022	—	`conf`	ARIA: Adversarially Robust Image Attribution for Content Provenance. CVPR Workshops Maksym Andriushchenko, Xiaoyang Rebecca Li, Geoffrey Oxholm, Thomas Gittings, Tu Bui, Nicolas Flammarion, John P. Collomosse
2022	J	`jnl`	ARIA: Adversarially Robust Image Attribution for Content Provenance. CoRR Maksym Andriushchenko, Xiaoyang Rebecca Li, Geoffrey Oxholm, Thomas Gittings, Tu Bui, Nicolas Flammarion, John P. Collomosse
2022	A	`conf`	On the effectiveness of adversarial training against common corruptions. UAI Klim Kireev, Maksym Andriushchenko, Nicolas Flammarion
2022	J	`jnl`	SGD with large step sizes learns sparse features. CoRR Maksym Andriushchenko, Aditya Varre, Loucas Pillaud-Vivien, Nicolas Flammarion
2022	A*	`conf`	Sparse-RS: A Versatile Framework for Query-Efficient Sparse Black-Box Adversarial Attacks. AAAI Francesco Croce, Maksym Andriushchenko, Naman D. Singh, Nicolas Flammarion, Matthias Hein
2022	A*	`conf`	Towards Understanding Sharpness-Aware Minimization. ICML Maksym Andriushchenko, Nicolas Flammarion
2022	J	`jnl`	Towards Understanding Sharpness-Aware Minimization. CoRR Maksym Andriushchenko, Nicolas Flammarion
2021	A*	`conf`	On the Stability of Fine-tuning BERT: Misconceptions, Explanations, and Strong Baselines. ICLR Marius Mosbach, Maksym Andriushchenko, Dietrich Klakow
2021	J	`jnl`	On the effectiveness of adversarial training against common corruptions. CoRR Klim Kireev, Maksym Andriushchenko, Nicolas Flammarion
2021	—	`conf`	RobustBench: a standardized adversarial robustness benchmark. NeurIPS Datasets and Benchmarks Francesco Croce, Maksym Andriushchenko, Vikash Sehwag, Edoardo Debenedetti, Nicolas Flammarion, Mung Chiang, Prateek Mittal, Matthias Hein
2020	J	`jnl`	On the Stability of Fine-tuning BERT: Misconceptions, Explanations, and Strong Baselines. CoRR Marius Mosbach, Maksym Andriushchenko, Dietrich Klakow
2020	J	`jnl`	RobustBench: a standardized adversarial robustness benchmark. CoRR Francesco Croce, Maksym Andriushchenko, Vikash Sehwag, Nicolas Flammarion, Mung Chiang, Prateek Mittal, Matthias Hein
2020	J	`jnl`	Sparse-RS: a versatile framework for query-efficient sparse black-box adversarial attacks. CoRR Francesco Croce, Maksym Andriushchenko, Naman D. Singh, Nicolas Flammarion, Matthias Hein
2020	—	`conf`	Square Attack: A Query-Efficient Black-Box Adversarial Attack via Random Search. ECCV (23) Maksym Andriushchenko, Francesco Croce, Nicolas Flammarion, Matthias Hein
2020	A*	`conf`	Understanding and Improving Fast Adversarial Training. NeurIPS Maksym Andriushchenko, Nicolas Flammarion
2020	J	`jnl`	Understanding and Improving Fast Adversarial Training. CoRR Maksym Andriushchenko, Nicolas Flammarion
2019	A	`conf`	Provable Robustness of ReLU networks via Maximization of Linear Regions. AISTATS Francesco Croce, Maksym Andriushchenko, Matthias Hein
2019	J	`jnl`	Provably Robust Boosted Decision Stumps and Trees against Adversarial Attacks. CoRR Maksym Andriushchenko, Matthias Hein
2019	A*	`conf`	Provably robust boosted decision stumps and trees against adversarial attacks. NeurIPS Maksym Andriushchenko, Matthias Hein
2019	J	`jnl`	Square Attack: a query-efficient black-box adversarial attack via random search. CoRR Maksym Andriushchenko, Francesco Croce, Nicolas Flammarion, Matthias Hein
2019	A*	`conf`	Why ReLU Networks Yield High-Confidence Predictions Far Away From the Training Data and How to Mitigate the Problem. CVPR Matthias Hein, Maksym Andriushchenko, Julian Bitterwolf
2019	—	`conf`	Why ReLU networks yield high-confidence predictions far away from the training data and how to mitigate the problem. CVPR Workshops Matthias Hein, Maksym Andriushchenko, Julian Bitterwolf
2018	J	`jnl`	Logit Pairing Methods Can Fool Gradient-Based Attacks. CoRR Marius Mosbach, Maksym Andriushchenko, Thomas Alexander Trost, Matthias Hein, Dietrich Klakow
2018	J	`jnl`	Provable Robustness of ReLU networks via Maximization of Linear Regions. CoRR Francesco Croce, Maksym Andriushchenko, Matthias Hein
2018	J	`jnl`	Why ReLU networks yield high-confidence predictions far away from the training data and how to mitigate the problem. CoRR Matthias Hein, Maksym Andriushchenko, Julian Bitterwolf
2017	—	`conf`	Formal Guarantees on the Robustness of a Classifier against Adversarial Manipulation. NIPS Matthias Hein, Maksym Andriushchenko
2017	J	`jnl`	Formal Guarantees on the Robustness of a Classifier against Adversarial Manipulation. CoRR Matthias Hein, Maksym Andriushchenko