Off the top of my head, typically 10-100mg of dye is required for 1 mol of silver (in the form of a silver-gelatin emulsion) The actual figure depends partly upon the size and surface area of the grains.
Therefore, it's not a simple matter to say x grams of dye will sensitize y number of plates.
When it comes to wet-plate, it may be a totally different number, or it might not work at all.
In the early days, plates were spectrally sensitized by soaking them in a dye solution.
"The Photographic Emulsion", Carrol & Hubbard, Ch11 touches upon the mechanisms and variables of dye sensitization.
A search on "dyes" and "spectral" on APUG should reveal greater detail.