Text Data For Training Large Language Models - Scope Ii