There are several papers on fluorescent gold clusters and perhaps even more on silver. For gold, there appears to be a very narrow size range (below 2nm) where the fluorescence is strong enough to be easily visible with the naked eye. Nevertheless, quantum yields tend to be low (approx. 1%). I am not aware of any reports on the fluorescence of ligand free gold clusters, so it is still hard to say if the fluorescence stems from ligand-core interactions or if it is entirely due to the properties of the metal cluster alone. In this sense, I believe the question is still open whether something like a "dye" is needed or not.