papers

2026

1. Temporal Sparse Autoencoders: Leveraging the Sequential Nature of Language for InterpretabilityOral

   ICLR

   previously Oral Presentation @ Workshop on Assessing World Models, ICML 2025

   U. Bhalla, A. Oesterling, C. Mayrink Verdun, H. Lakkaraju, F. Calmon

2. GradPCA: Leveraging NTK Alignment for Reliable Out-of-Distribution Detection

   ICLR

   M. Seleznova, H. Chou, C. Mayrink Verdun, G. Kutyniok

3. Evaluating large language models’ ability to automate spear phishing

   Expert Systems with Applications

   F. Heiding, S. Lermen, A. Kao, C. Mayrink Verdun, B. Schneier, A. Vishwanath

4. Reading Between the Dots: Decoding Hidden Computation across Filler Tokens

   Submitted

   K. Brauer, C. Mayrink Verdun, S. Marks

5. An Instrument to Evaluate Governance Proposals: AI Policy Analysis at Scale

   Preprint

   P. Carvão, I. Adler, C. Mayrink Verdun, J. Zhou

   Featured at Tech Policy Press

2025

1. Multi-Group Proportional Representations for Text-to-Image Models

   CVPR

   previously Algorithmic Fairness through the lens of Metrics and Evaluation Workshop, NeurIPS 2024

   S. Jung, A. Oesterling, C. Mayrink Verdun, S. Vithana, T. Moon, F. Calmon

2. Get rid of your constraints and reparametrize: A study in NNLS and implicit bias

   AISTATS

   H. Chou, J. Maly, C. Mayrink Verdun, B. Costa, H. Mirandola

3. Inference-Time Reward Hacking in Large Language ModelsSpotlight

   NeurIPS

   previously Workshop on Models of Human Feedback for AI Alignment, ICML 2025

   H. Khalaf, C. Mayrink Verdun, A. Oesterling, H. Lakkaraju, F. Calmon

4. ProofCompass: Enhancing Specialized Provers with LLM Guidance

   ICML

   N. Wischermann, C. Mayrink Verdun, G. Poesia, F. Noseda

5. HeavyWater and SimplexWater: Watermarking Low-Entropy Text Distributions

   NeurIPS

   D. Tsur, C. Long, C. Mayrink Verdun, H. Hsu, C. Chen, H. Permuter, S. Vithana, F. Calmon

6. AI Alignment at Your DiscretionBEST PAPER · NENLP 2025

   FAccT

   M. Buyl, H. Khalaf, C. Mayrink Verdun, L. Paes, C. Machado, F. Calmon

7. Optimized Couplings for Watermarking Large Language Models

   ISIT

   C. Long, D. Tsur, C. Verdun, H. Hsu, H. Permuter, F. Calmon

8. Soft Best-of-n Sampling for Model Alignment

   ISIT

   C. Mayrink Verdun, A. Oesterling, H. Lakkaraju, F. Calmon

2024

1. Non-Asymptotic Uncertainty Quantification in High-Dimensional LearningSpotlight

   NeurIPS

   F. Hoppe, C. Mayrink Verdun, H. Laus, F. Krahmer, H. Rauhut

2. Measuring progress in dictionary learning for language model interpretability with board game models

   NeurIPS

   previously Oral Presentation @ Mechanistic Interpretability Workshop, ICML 2024

   A. Karvonen, B. Wright, C. Rager, R. Angell, J. Brinkmann, L. Smith, C. Mayrink Verdun, D. Bau, S. Marks

3. Multi-Group Proportional Representation

   NeurIPS

   A. Oesterling, C. Mayrink Verdun, C. Long, A. Glynn, L. Paes, S. Vithana, M. Cardone, F. Calmon

4. Imaging with Confidence: Uncertainty Quantification for High-dimensional Undersampled MR Images

   ECCV

   F. Hoppe, C. Mayrink Verdun, H. Laus, S. Endt, M. Menzel, F. Krahmer, H. Rauhut

5. Fast, blind, and accurate: Tunning-free sparse regression with global linear convergence

   COLT

   C. Mayrink Verdun, O. Melnyk, F. Krahmer, P. Jung

6. With or Without Replacement? Improving Confidence in Fourier Imaging

   COSERA

   F. Hoppe, C. Mayrink Verdun, F. Krahmer, M. Menzel, H. Rauhut

7. AI Technologies: Algorithmic Monocultures, Arbitrariness, and Global Divides

   T20 conference

   C. Mayrink Verdun, C. Tsuzuki, C. Machado, J. Gomez, L. Monteiro Paes, F. Calmon

2023

1. Uncertainty quantification for learned ISTA

   MLSP

   F. Hoppe, C. Mayrink Verdun, H. Laus, F. Krahmer, H. Rauhut

2. In-vivo myelin water quantification using diffusion-relaxation correlation MRI: a comparison of 1D and 2D methods

   Applied Magnetic Resonance

   S. Endt, M. Engel, E. Naldi, R. Assereto, M. Molendowska, L. Müller, C. Mayrink Verdun, C. Pirkl, M. Palombo, D. Jones, M. Menzel

3. Sampling Strategies for Compressive Imaging Under Statistical Noise

   SampTA 2023

   F. Hoppe, F. Krahmer, C. Mayrink Verdun, M. Menzel, H. Rauhut

4. High-Dimensional Confidence Regions in Sparse MRIBest Student Paper

   ICASSP

   F. Hoppe, F. Krahmer, C. Mayrink Verdun, M. Menzel, H. Rauhut

2022

1. Uncertainty quantification for sparse Fourier recovery

   arXiv preprint arXiv:2212.14864

   F. Hoppe, F. Krahmer, C. Mayrink Verdun, M. Menzel, H. Rauhut

2. Non-negative least squares via overparametrization

   arXiv preprint arXiv:2207.08437

   H. Chou, J. Maly, C. Mayrink Verdun

2021

1. Unmixing tissue compartments via deep learning T1-T2-relaxation correlation imaging

   17th International Symposium on Medical Information Processing and Analysis

   S. Endt, C. Pirkl, C. Mayrink Verdun, B. Menze, M. Menzel

2. A scalable second order method for ill-conditioned matrix completion from few samplesSpotlight

   ICML

   C. Kümmerle, C. Mayrink Verdun

3. Group testing for SARS-CoV-2 allows for up to 10-fold efficiency increase across realistic scenarios and testing strategies

   Frontiers in Public Health

   C. Mayrink Verdun, T. Fuchs, P. Harar, D. Elbrächter, D. Fischer, J. Berner, P. Grohs, F. Theis, F. Krahmer

   Highlighted by David Donoho at SIAM Mathematics of Data Science Distinguished Lecture

4. Iteratively reweighted least squares for basis pursuit with global linear convergence rateSpotlight

   NeurIPS

   C. Kümmerle, C. Mayrink Verdun, D. Stöger

2020

1. Escaping Saddle Points in Ill-Conditioned Matrix Completion with a Scalable Second Order Method

   ICML

   C. Kümmerle, C. Mayrink Verdun

2019

1. Completion of structured low-rank matrices via iteratively reweighted least squares

   SampTA

   C. Kümmerle, C. Mayrink Verdun