Prompting pre-trained language models has achieved impressive performance on
various NLP tasks, especially in low data regimes. Despite the success of
prompting in monolingual settings, applying prompt-based methods in
multilingual scenarios has been limited to a narrow set of tasks, d