Scholar
Patrik Wolf
Google Scholar ID: vx3ng84AAAAJ
PhD student, ETH Zurich
Citations & Impact
Citations
0
H-index
0
i10-index
0
Publications
1
Co-authors
2
list available
Publications
Resume (English only)
Academic Achievements
- Paper 'Specialization after Generalization: Towards Understanding Test-Time Training in Foundation Models' presented as an oral presentation at NeurIPS 2025 CCFM Workshop; Paper 'Understanding Gradient Flow Dynamics for Matrix Factorization Problems' awarded an ETH Medal.
Research Experience
- Conducted research on gradient flow dynamics for matrix factorization problems at ETH Zürich and Harvard (February 2024 - September 2024), and a project on specialization after generalization through test-time training in foundation models.
Education
- Doctoral researcher at ETH Zürich, advised by Andreas Krause; Researcher at MPI-IS Tübingen, advised by Celestine Mendler-Dünner.
Background
- Research Interests: Test-time training and preference alignment, often in the context of large language models; aims to develop a theoretical understanding of these techniques and explore their potential to improve model performance.
