updated docs of GP formula fn to include more details on hyperparameters

[jr-leary7]

in response to #1725, i added some further documentation on the hyperparameters estimated by a GP, specifically the length-scale $\ell$ and the marginal standard deviation $\sigma$, including general guidelines on how to interpret each hyperparameter.

what remains is to add the name, support, and simple interpretation of the hyperparameter $z$ to the relevant table in the roxygen2 doc section of gp(), as this bit is unclear to me. i left the two cells in the table that i was unable to fill in marked with ???. i would also appreciate if someone gave a quick look over the information i've provided to ensure correctness, though i mostly took the info straight from the MC-Stan docs and forums.

[paul-buerkner]

Great, thanks! I will take a closer look and merge it before the next release.