About Me

Hi! I’m Katharina, a PhD student in the Computational Linguistics Group of the Department of Language Science and Technology at Saarland University. I am supervised by Prof. Alexander Koller and Prof. Jörg Hoffmann.

I’m working in the A7 project of the Collaborative Reserach Center on Information Density and Linguistic Encoding (SFB 1102) and I’m associated with the Research Training Group on Neuroexplicit Models.

The main topic of my PhD is on exploring the capabilities and limitations of using LLMs for complex reasoning tasks, in particular on classical planning tasks. I’m particularly interested in integrating LLM-based methods with symbolic methods that obtain approaches that reliably solving planning tasks and generalize to tasks of different complexity. Currently, I’m exploring the potential of LLMs to generate programs that solve planning tasks and generalize within a domain.

I’m a co-organizer of the Workshop on Planning in the Era of LLMs at ICAPS 2026 and ICML 2026

My research interests include

• Large Language Model for Planning and Reasoning
• (Length) generalization in the context of LLMs and planning
• LLM-based code generation
• Combining LLMs with symbolic approaches to increase reliability and obtain guarantees
• Learning from and adaption through interactions

Publications

Improved Generalized Planning with LLMs through Strategy Refinement and Reflection.
Katharina Stein, Nils Hodel, Daniel Fišer, Jörg Hoffmann, Michael Katz, and Alexander Koller
@ICAPS 2026

On the Ability of Transformers to Verify Plans
Yash Sarrof, Yupei Du, Katharina Stein, Alexander Koller, Sylvie Thiébaux, Michael Hahn
@ICML 2026

Automating the Generation of Prompts for LLM-based Action Choice in PDDL Planning
Katharina Stein, Dan Fišer, Jörg Hoffmann and Alexander Koller
@ICAPS 2025

See here for full list