We can't observe the results of these mergers because they are either isolated black holes or they are isolated, massive neutron stars.
In neither case is there anything to observe. Compact objects that are not accreting significant amounts of material from a binary companion are either invisible (black holes) or extremely faint (tiny neutron stars).
It is possible I suppose that the merger could happen in a hierarchical triple system (a close binary plus a much wider third body), but then the resulting binary would be too wide for significant mass transfer and accretion to occur. So again, the result is a nearly invisible merger remnant. On the other hand it might be possible for the post-merger object to get closer to or even capture a new companion if it is in a dense cluster (an example may have been found by Barr et al. 2024).
There is also the probability that any newly formed massive neutron star is hot for some time after the merger. But the cooling timescales are less than a million years or so - a blink of the eye in cosmic terms - and even a very hot neutron star would need to be close to us (within a few hundred pc) to be observable. Even the one thousand-year old Crab pulsar has an optical absolute magnitude of just 5. Any pulsar-like activity is also expected to disappear on similar timescales.
Note that the "mass gap" is just a (relative) absence of observed compact objects in the range $2.5-5 M_\odot$. This may be because those objects are rare (and neutron star mergers are rare - see below) or it may be because they are difficult to observe - see the discussion in https://astronomy.stackexchange.com/a/8280/2531 . A review of the available evidence for candidate objects populating the mass gap by Da Sa et al. (2022) concludes that there isn't a complete absence of objects in this mass range and neutron star mergers is one way in which the gap might be populated.
Addition:
The neutron star merger rate estimated from the LIGO gravitational wave data is $10^2-4\times 10^3$ (Gpc)$^{-3}$/year (Abbott et al. 2019). The rate of local kilonovae, mostly thought to be caused by neutron star mergers, is $<9\times 10^2$ (Gpc)$^{-3}$/year (Andreoni et al. 2021).
The rate at which black holes are born can be estimated from the star formation rate in the local universe of $2.5\times 10^7 M_\odot$ (Gpc)$^{-3}$/year (Bothwell et al. 2011). Then, assuming a Salpeter-like stellar initial mass function (see for example here) and assuming that all stars with initial mass $>20 M_\odot$ become black holes, we can estimate that the black hole birth rate is about $6\times 10^4$ (Gpc)$^{-3}$/year.
Thus it seems unlikely that neutron star mergers, even if they all result in black holes (some will result in stable neutron stars - Piro et al. 2017), can provide more than a few per cent contribution to the overall black hole population, and as I said above, most will be isolated and unobservable.
