Abstractions can contain entities and other abstractions. When an abstraction is activated, only those entities and abstractions contained within the abstraction will be shown on the model diagram. Only those relationships that link the entities within the abstraction are shown. Refer to 8.3.11 Hiding Complexity.
In addition, the Build Model function is restricted to those entities within the abstraction, enabling you to build discreet parts of the model. Changes to entities and relationships within the abstraction are still propagated to the entire model, however, and can affect entities that are not included in the abstraction.