The size of the VLP and the ordinary viral particle should be similar, are you sure you don't have any artifacts in your expression system (e.g. fragment size, start codon position, etc)? You can also try another expression system.
When I was producing influenza VLPs using drosophila S2 cell line, the nucleocapsid protein self-assembled particles without any surface glycoprotein have more irregular shapes and sizes compared with particles with HA and NA. The NP alone ones are quite often slightly smaller, but occasionally really bigger.
I'm not sure if it's a expression system thing, as the difference was not so obvious in 293FT expressed VLPs. But it's just my vague feeling. Bare in mind, it was gag protein from HIV instead of NP from influenza used as nucleocapsid in S2 and 293.
Whether a VLP's morphology matches that of the corresponding virus can vary a lot- with flu many of the VLPs are irregular in size and shape (unlike influenza virions). Ideally you can also confirm with immunogold that the smaller "VLP" structures with NP you are identifying definitely have NP in them (as cells can excrete all kinds of membrane bound vesicles, many of which look similar to VLPs). Without the characteristic "fringe" seen with the presence of viral glycoproteins it can be quite tricky to tell a VLP from an exosome.