The objective of the present paper is to show that predictions of the conventional strain gradient theories do not coincide with some general physical expectations when large strains and geometry changes should be considered. As an alternative, it is proposed to use strain rate gradient theories of plasticity. One possible theory of this type is formulated as a formal modification of a strain gradient theory of plasticity. The problem of hollow sphere expansion at large strains is solved for both the strain gradient and strain rate gradient theories of plasticity. Comparison of these solutions reveals advantages of the strain rate gradient theory of plasticity for a class of problems.