Modern language models can imitate complex patterns through few-shot
lea...
Transparency methods such as model visualizations provide information th...
To understand neural network behavior, recent works quantitatively compa...
In this contribution, we propose a new computationally efficient method ...
share