Research Interests
Differentiable Optimization, Safety of LLMs, Reinforcement Learning, Differentiable Simulation
Short CV
- 2022-present: PhD student, IMPRS-IS with MPI-IS and Uni Tübingen
- 2021-2022: M.Sc. Physics, Uni Tübingen
- 2018-2021: B.Sc. Computer Science, Uni Tübingen
- 2015-2018: B.Sc. Physics, Uni Tübingen
Selected Publications
- Paulus A*, Zharmgambetov A*, Guo C, Amos B, Tian Y: AdvPrompter: Fast Adaptive Adversarial Prompting for LLMs. Currently under review
- Paulus A, Martius G, Musil V: LPGD: A General Framework for Backpropagation through Embedded Optimization Layers. Published at ICML 2024
- Vlastelica M*, Paulus A*, Musil V, Martius G, Rolinek M: Differentiation of Blackbox Combinatorial Solvers. Published at ICLR 2020