The reason for separation is to get more than 1 sec of arc difference; at that range, most computers can get a fix based upon two directional lines and find te point source.
With a directional (azimuth and elev) sensor that provides magnitudes, you only need 2 separate fixes to determine range and location.
With magnitudes only, you need no more than 4; a triangle on a plane, and one above plane. With this you can do the trig to find the point source.
Current S&R using ELT fixes is usually done by using a two-dimentional DF reciever, man portable (L-Per). Two fixes from a distance of a mile apart will provide a good bearing plot on anything within 3-4 miles; you don't even need the magnitude reading for determining distance. (Although, if the ELT is at full charge, it's usually indicative.) The L-Per I used as a cadet in the late 80's weighed less than 2KG, used a dual antenna with a fixed 2ft separation. In a few odd drills, we turned it sideways to gain an elevation fix. (In one, it was sitting on top of a building).
Neutrino sensors are directional, in that the smashed particle goes in the direction smashed, and triggers a descending luminesscence trail away from source. (The exact internals of HOW it luminesces are immaterial for the most part; it is enough to know that it does.)
A few hundred meters separation (easily done on a battleship) gives reliable fixes for a goodly distance.