Ah one of my old posts has been resurrected. Ill start by answering the non-controversial questions. As to how the detectors work. I said neutrinos interact very weakly, not, they don't interact. So to detect them you need a huge detector and a lot of neutrinos. The most famous detector is Superkamiokande in Japan, which is filled with 50'000 tonnes of water. There are other designs, Iron is another common material, denser than water so you don't need as much, although its a pain to see anything with photo sensors as Iron is not transparent to light. You get round this by having alternating layers of Iron and plastic.
As to where the experiments are located, in vague chronological order;
K2K - source: KEK, Tsukuba, Japan detector: Superkamiokande, Japan
KamLAND - source: nuclear reactors all over Japan detector: Same place as Superkamiokande, Japan
MINOS - source: Fermilab, Batavia, Illinois, US detector: Soudan, Minnesota, US
CNGS - source: CERN-SPS, Geneva, Switzerland detector: Gran Sasso, Italy (near where the earthquake was)
T2K - source: Tokai, Japan detector: Superkamiokande, Japan (Construction nearly completed, beginning data runs in December 09/Jan 10)
NOVA - source: Fermilab, Batavia, Illinois, US detector: Ash River (I think), Minnesota, US
A plethora of short baseline experments also exist but are not really relevant here. The experiments above have a baseline between 200km and 1000km, roughly. This corresponds to the beam passing between about 1 and 10 miles deep through the crust, if you do the math assuming the Earth that all bar a few of people accept. The mother of all long baseline neutrino experiments, the neutrino factor is being designed to have a baseline of about 7000km hopefully construction will begin when someone figure out how to make a beamline going down in to the ground at a steep angle. Current long baseline experiments only need to be a few degrees from the horizontal.