Unveiling the Potential of Superexpressive Networks in Implicit Neural Representations