Some reasoning methods for Defeasible Description Logics (DDLs) suffer from quantification neglect (QN) as they omit un-defeated information for quantified objects. Reasoning in defeasible EL_⊥ based on so-called typicality models (TMs), which extend canonical models of classical EL_⊥, can alleviate QN. The DDL ELI_⊥ extends EL_⊥ by inverse roles, i.e., a limited form of value restriction. Extending TMs to inverse roles is challenging due to their interaction with existential restrictions. In this paper, we develop TMs for ELI_⊥ for 4 different semantics reliant on rational and relevant closure. Our computation methods for those TMs are effective decision procedures for subsumption in defeasible ELI_⊥ and the stronger forms of TMs can mitigate QN.