The top comment there explains a good use-case for this chip.
[static html page saying "no."]
This reminds me a bit of the XKCD question about how close you need to be to a supernova to die from neutrinos: inside it.
So testing and calibrating it would consist of exposing it to a source of known strength and making sure it trips. For this kind of setup, the source doesn't have to be an actual nuclear weapon going off or whatever target event, it'll trip for any sort of gamma or neutron source (presumably, it's light on details and I haven't ever seen one of these in the wild). There are a wide range of suitable industrial sources available commercially.
Edit: you can safely assume that the manufacturer provides a detailed calibration procedure wherever the more complete documentation is. Probably specifies what test sources to use and how to correlate them with different types of target events, which they determined through simulation and laboratory measurements. Also worth noting that this thing is very likely entirely analog, given its intended use.
I suspect that or a similar methodology is in use.