Invited talks
2025
-
On the Generalization Ability of Transformers
Fudan University, Nov 2025 -
Language models can learn implicit multi-hop reasoning, but only if they have lots of training data
Max Planck Institute for Software Systems (MPI-SWS), Jul 2025
2024
-
Simple and effective data augmentation for compositional generalization
DFKI, Multilinguality and Language Technology Lab, Apr 2024