You should verify that atom distances (like C-H bond lengths) were correct, and the coordinates are in A instead of something else like Bohr. (If it is not correct, which is likely, check whether all distances are enlarged by a factor of 1.89). There might be an issue with the unit during topmod processing or the software you calculate the electronic structure.

And, use Multiwfn anyway.

Using Multiwfn to perform ELF basin analysis is much better than Topmod, which is out-of-date, not flexible and not as powerful as Multiwfn. Multiwfn itself is also able to directly visualize the located attractors. See Section 4.17.2 of Multiwfn manual for example.

Recent example:
Computational and Theoretical Chemistry 1154 (2019) 17–25
and
DOI: 10.1177/1468678319825742