MV

Martin Vechev

0000-0002-0054-9568

cs.LGcs.AIcs.CRcs.CLcs.PLcs.SEstat.MLcs.CVcs.DCcs.CY

On Valency

published · living versions

Can Open Be Too Open? On the Security and Trustworthiness of Fully Transparent LLMs

with Sheila Zingg, Florian Tramèr, Srdjan Čapkun

1 version

W_t69j7389v1 · current

End-to-End Neural Modeling of EM Response and Design Performance for Free-Form RFIC Passives

W_6cc2ptuxv1

PROBE: Frontier Coding Agents Find Different Bugs Than Maintainers Fix

W_6v7x34v3v1

Preprints & journals

133 papers in the corpus · 2016–2026

TestSuiteBench: Coding Agents Struggle to Generate Strong Test Suites from Specifications

Kazuki Egashira, Niels Mundler-Sasahara, Mark Niklas Muller et al.

0 citations

2610.13065v12026arXiv

AutoMark: Enabling Autoresearch to Discover Better LLM Watermarks

Thibaud Gloaguen, Robin Staab, Martin Vechev

0 citations

2609.37310v12026arXiv

Leveraging Instruction Tuning and Merging for Reasoning Model Adaptation

Yu-Du Feng, Niels Mundler-Sasahara, Mark Vero et al.

0 citations

2607.14895v22026arXiv

Generative Compilation: On-the-Fly Compiler Feedback as AI Generates Code

Niels Mundler-Sasahara, Hristo Venev, Dawn Song et al.

0 citations

2607.13921v22026arXiv

SABER-Math: Automated Benchmark for Information Retrieval Evaluation in Mathematics

Nikolay Georgiev, Maria Drencheva, Kseniia Ibragimova et al.

0 citations

2606.29894v22026arXiv

TIGER: Inverting Transformer Gradients via Embedding-Subspace Distance Optimization

William Kalikman, Ivo Petrov, Dimitar I. Dimitrov et al.

0 citations

2606.18312v12026arXiv

Optimizing the Cost-Quality Tradeoff of Agentic Theorem Provers in Lean

K'ari Rognvaldsson, Chenhao Sun, Jasper Dekoninck et al.

0 citations

2606.04883v32026arXiv

Learning from Saturated Data: Signals Beyond Correctness for LLM Training

Hanno Hiss, Jasper Dekoninck, Martin Vechev

0 citations

2606.01436v12026arXiv

Marking the Wrong Symptoms: Evaluating LLM Watermarks in Medical Texts

Melanie Rieff, Robin Staab, Thibaud Gloaguen et al.

0 citations

2607.20462v12026arXiv

Widening the Gap: Exploiting LLM Quantization via Outlier Injection

Xiaohua Zhan, Kazuki Egashira, Robin Staab et al.

0 citations

2605.15152v22026arXiv

Every Bit, Everywhere, All at Once: A Binomial Multibit LLM Watermark

Thibaud Gloaguen, Robin Staab, Mark Vero et al.

0 citations

2605.11653v22026arXiv

Making Open-Source Text LLM Watermarks Durable Against Merging

Luisa Scharff, Thibaud Gloaguen, Robin Staab et al.

0 citations

2607.20435v12026arXiv

Coding Agents Don't Know When to Act

Thibaud Gloaguen, Niels Mundler, Mark Muller et al.

0 citations

2605.07769v12026arXiv

Beyond Benchmarks: MathArena as an Evaluation Platform for Mathematics with LLMs

Jasper Dekoninck, Nikola Jovanovi'c, Tim Gehrunger et al.

0 citations

2605.00674v22026arXiv

Delay, Plateau, or Collapse: Evaluating the Impact of Systematic Verification Error on RLVR

Kazuki Egashira, Mark Vero, Jasper Dekoninck et al.

0 citations

2605.02909v22026arXiv

SecPI: Secure Code Generation with Reasoning Models via Security Reasoning Internalization

Hao Wang, Niels Mundler, Mark Vero et al.

0 citations

2604.03587v12026arXiv

CodeTaste: Can LLMs Generate Human-Level Code Refactorings?

Alex Thillen, Niels Mundler, Veselin Raychev et al.

0 citations

2603.04177v22026arXiv

Recovered in Translation: Efficient Pipeline for Automated Translation of Benchmarks and Datasets

Hanna Yukhymenko, Anton Alexandrov, Martin Vechev

0 citations

2602.22207v12026arXiv

Evaluating AGENTS.md: Are Repository-Level Context Files Helpful for Coding Agents?

Thibaud Gloaguen, Niels Mundler-Sasahara, Mark Niklas Muller et al.

1 citation

2602.11988v32026arXiv

A Unified Framework for LLM Watermarks

Thibaud Gloaguen, Robin Staab, Nikola Jovanovi'c et al.

0 citations

2602.06754v12026arXiv

Learning Compact Boolean Networks

Shengpu Wang, Yuhao Mao, Yani Zhang et al.

0 citations

2602.05830v22026arXiv

AutoBaxBuilder: Bootstrapping Code Security Benchmarking

Tobias von Arx, Niels Mundler, Mark Vero et al.

0 citations

2512.21132v22025arXiv

SPEAR++: Scaling Gradient Inversion via Sparsely-Used Dictionary Learning

Alexander Bakarsky, Dimitar I. Dimitrov, Maximilian Baader et al.

0 citations

2510.24200v12025arXiv

Pay Attention to the Triggers: Constructing Backdoors That Survive Distillation

Giovanni De Muri, Mark Vero, Robin Staab et al.

0 citations

2510.18541v12025arXiv

Adaptive Generation of Bias-Eliciting Questions for LLMs

Robin Staab, Jasper Dekoninck, Maximilian Baader et al.

0 citations

2510.12857v22025arXiv

Fewer Weights, More Problems: A Practical Attack on LLM Pruning

Kazuki Egashira, Robin Staab, Thibaud Gloaguen et al.

0 citations

2510.07985v32025arXiv

BrokenMath: A Benchmark for Sycophancy in Theorem Proving with LLMs

Ivo Petrov, Jasper Dekoninck, Martin Vechev

3 citations

2510.04721v12025arXiv

Watermarking Diffusion Language Models

Thibaud Gloaguen, Robin Staab, Nikola Jovanovi'c et al.

0 citations

2509.24368v32025arXiv

SoK: Data Minimization in Machine Learning

Robin Staab, Nikola Jovanovi'c, Kimberly Mai et al.

1 citation

2508.10836v22025arXiv

Constrained Decoding of Diffusion LLMs with Context-Free Grammars

Niels Mundler, Jasper Dekoninck, Martin Vechev

0 citations

2508.10111v12025arXiv

The Open Proof Corpus: A Large-Scale Study of LLM-Generated Mathematical Proofs

Jasper Dekoninck, Ivo Petrov, Kristian Minchev et al.

1 citation

2506.21621v22025arXiv

Watermarking Autoregressive Image Generation

Nikola Jovanovi'c, Ismail Labiad, Tom'as Soucek et al.

1 citation

2506.16349v22025arXiv

MathArena: Evaluating LLMs on Uncontaminated Math Competitions

Mislav Balunovi'c, Jasper Dekoninck, Ivo Petrov et al.

5 citations

2505.23281v32025arXiv

Mind the Gap: A Practical Attack on GGUF Quantization

Kazuki Egashira, Robin Staab, Mark Vero et al.

1 citation

2505.23786v32025arXiv

Watch your steps: Dormant Adversarial Behaviors that Activate upon LLM Finetuning

Thibaud Gloaguen, Mark Vero, Robin Staab et al.

0 citations

2505.16567v42025arXiv

LLM Fingerprinting via Semantically Conditioned Watermarks

Thibaud Gloaguen, Robin Staab, Nikola Jovanovi'c et al.

0 citations

2505.16723v32025arXiv

MixAT: Combining Continuous and Discrete Adversarial Training for LLMs

Csaba D'ek'any, Stefan Balauca, Robin Staab et al.

0 citations

2505.16947v22025arXiv

Type-Constrained Code Generation with Language Models

Niels Mundler, Jingxuan He, Hao Wang et al.

13 citations

2504.09246v22025arXiv

Proof or Bluff? Evaluating LLMs on 2025 USA Math Olympiad

Ivo Petrov, Jasper Dekoninck, Lyuben Baltadzhiev et al.

1 citation

2503.21934v52025arXiv

Automated Benchmark Generation for Repository-Level Coding Tasks

Konstantinos Vergopoulos, Mark Niklas Muller, Martin Vechev

0 citations

2503.07701v12025arXiv

ToolFuzz -- Automated Agent Tool Testing

Ivan Milev, Mislav Balunovi'c, Maximilian Baader et al.

1 citation

2503.04479v32025arXiv

GRAIN: Exact Graph Reconstruction from Gradients

Maria Drencheva, Ivo Petrov, Maximilian Baader et al.

0 citations

2503.01838v12025arXiv

BaxBench: Can LLMs Generate Correct and Secure Backends?

Mark Vero, Niels Mundler, Victor Chibotaru et al.

2 citations

2502.11844v32025arXiv

MathConstruct: Challenging LLM Reasoning with Constructive Proofs

Mislav Balunovi'c, Jasper Dekoninck, Nikola Jovanovi'c et al.

0 citations

2502.10197v22025arXiv

Towards Watermarking of Open-Source LLMs

Thibaud Gloaguen, Nikola Jovanovi'c, Robin Staab et al.

0 citations

2502.10525v12025arXiv

BgGPT 1.0: Extending English-centric LLMs to other languages

Anton Alexandrov, Veselin Raychev, Dimitar I. Dimitrov et al.

0 citations

2412.10893v12024arXiv

COMPL-AI Framework: A Technical Interpretation and LLM Benchmarking Suite for the EU Artificial Intelligence Act

Philipp Guldimann, Alexander Spiridonov, Robin Staab et al.

5 citations

2410.07959v22024arXiv

Average Certified Radius is a Poor Metric for Randomized Smoothing

Chenhao Sun, Yuhao Mao, Mark Niklas Muller et al.

0 citations

2410.06895v42024arXiv

Ward: Provable RAG Dataset Inference via LLM Watermarks

Nikola Jovanovi'c, Robin Staab, Maximilian Baader et al.

0 citations

2410.03537v22024arXiv

AlphaIntegrator: Transformer Action Search for Symbolic Integration Proofs

Mert Unsal, Timon Gehr, Martin Vechev

0 citations

2410.02666v12024arXiv

Career total: 392 works. 133 are in this corpus.Showing the 50 most recent.

Profile built from the corpus for this byline.

392

works

11.8K

citations

61

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.