Project

Cracking the Multilingual Code: Exploring Multilinguality in LLMs through Mechanistic Interpretability