This paper presents the shake table testing and finite element (FE) modeling of a modular prefabricated concrete bridge‐like specimen. The specimen comprised four equal‐height cylindrical reinforced concrete (RC) columns capped with an RC slab. The structural connections were non‐monolithic. Hence, controlled relative motion of the members, including rocking (uplift) of the piers, was allowed. The columns were connected to the slab with stiff tendons that provided positive post‐uplift stiffness. The specimen was subjected to 184 triaxial shake table tests, so that a statistical validation of numerical models can be performed. Subsequently, a detailed three‐dimensional FE model of the bridge was developed. The objectives of the present study were to: i) investigate the shake table response of a modular bridge with positive post‐uplift stiffness under multiple ground motions, ii) develop an FE model of the proposed structural system, iii) investigate the influence of geometrical imperfections on rocking bridges, and iv) evaluate the efficiency of using additional dissipative rebars. After being subjected to 184 shake table tests, the specimen showed zero damage, moderate displacements and tendon forces (TFs), low slab torsion, and zero residual displacements. The shake table tests were practically repeatable. The proposed FE model accurately captured the experimental results. Geometrical imperfections heavily affect the response of negative stiffness systems. However, they have a marginal influence on positive stiffness systems. When comparing systems with equivalent uplift resistance and post‐uplift stiffness, the use of additional dissipative rebars results in lower slab torsion and TFs, provided that the rebars do not fracture.