Maksym Andriushchenko
Scholar

Maksym Andriushchenko

Google Scholar ID: ZNtuJYoAAAAJ
ELLIS Institute Tübingen & Max Planck Institute for Intelligent Systems
AI SafetyAI AlignmentLLMsLLM agents
Homepage↗Google Scholar↗
Citations & Impact
All-time
Citations
7,191
 
H-index
24
 
i10-index
27
 
Publications
20
 
Co-authors
17
list available
Contact
Emailmaksym@andriushchenko.meCVOpen ↗TwitterOpen ↗GitHubOpen ↗LinkedInOpen ↗
Publications
32 items
AgentTime: Can Agents Estimate and Control Their Own Runtime?
2026
Cited
0
Alignment via Training Against Probes Without Losing Monitorability
2026
Cited
0
LLM Agents Can Easily Tamper With Their Own Traces
2026
Cited
0
Instrumental Monitor Evasion Emerges Under Ordinary Task Pressure
2026
Cited
0
Stealing Reasoning Traces from Proprietary LLM APIs
2026
Cited
0
ResearchArena: Evaluating Sabotage and Monitoring in Automated AI R&D
2026
Cited
0
What Shapes Emergent Misalignment? Insights from Training Dynamics, Model Priors, and Data
2026
Cited
0
Decomposing and Measuring Evaluation Awareness
2026
Cited
0
Co-authors
17 total
Nicolas Flammarion
Nicolas Flammarion
EPFL
Matthias Hein
Matthias Hein
Professor of Computer Science, University of Tübingen
Francesco Croce
Francesco Croce
EPFL
Edoardo Debenedetti
Edoardo Debenedetti
ETH Zurich
Matt Fredrikson
Matt Fredrikson
Carnegie Mellon University
Marius Mosbach
Marius Mosbach
Mila - Quebec AI Institute, McGill University
Zico Kolter
Zico Kolter
Carnegie Mellon University
Andy Zou
Andy Zou
PhD Student, Carnegie Mellon University