computational chemistry - Why the optimization in gaussian shows no bond with some atoms in diborane?

I have optimized the structure of diborane with Gaussian at the B3LYP/6-31+g(d,p) level and when check it on Gaussview I found that there are two hydrogens with no bond to boron atoms (like free atoms) and the two B atoms are bonded. Frequency calculations gives zero imaginary frequencies and when I compute properties (like ionization potential) the values are good approximations.

I wonder if the problem is with the calculation or only the graphical interface of Gaussview doesn't recognize it as an actual bond.

Also, the bond distances are in accordance with the literature (~1.3 Angstroms for H shared between the borons and ~1.19 Angstroms for the others).