Suppose there's an interaction AB -> CD, that goes via a virtual W boson, my question is, will the virtual boson leaves a track in the detector even though it's virtual ?
Just to clear up some doubts... If it's a virtual particle produced in the s-channel say, the whole AB->CD interaction will be happening at one contact point, ie, four tracks (due to ABCD), unlike say AB->2 photons, and this photon flies off some distance before it interacts with the detector to give electron-positron pair, which in this case, there is two contact point, first ABphoton and the other point, photon-electron-positron.No is the short answer to your question.
Think about what virtual means. A particle is virtual i.e "off mass shell" (violating energy conservation) for a time permitted by the uncertainty principle. Hence by definition it's impossible to "resolve" a virtual particle. Hence the name virtual.