Simon Ostermann

Simon Ostermann

NLP & AI Researcher @ DFKI. Mechanistic interpretability and multilingual NLP, with a focus on low-resource languages.

When Tokenization is Secretly Output Supervision

Published in Accepted at the 2026 Conference on Empirical Methods in Natural Language Processing (Main), 2026

Access paper here

Recommended citation: Tanja Baeumel, Josef van Genabith, and Simon Ostermann. (2026). "When Tokenization is Secretly Output Supervision." Accepted at the 2026 Conference on Empirical Methods in Natural Language Processing (Main).
Download Paper