Ben Bales

What does a vector sigma look like here? Rasmussen and Williams says the dot product kernel is: k(x1, x2) = sigma^2 + x1^T * x2 How does a sigma that's...

Aah, got it, thanks both

Actually question @avehtari, should I implement this just as: k(x1, x2) = x1^T diag(sigma^2) x2? Or should I add another argument to do: k(x1, x2) = alpha_0 + x1^T diag(sigma^2)...

@betanalpha I don't have any plans to make this efficient. Just being a code-robot on this one. But we should still get the interface right so it doesn't have to...

(they're all easy enough to implement)

Just to clarify, when the input is a std::vector of Eigen::VectorXds, I'm only allowing the vector sigma. I'm not allowing a scalar in that case. This is consistent with how...

We could try to make this easy like the \_lpdf functions (where in a lot of cases vectors and arrays can get swapped around), but to make things consistent then...

Unfortunately I think this is limited by my actual laziness, not stan3 :D. I'll go with @betanalpha's signature since it's more general. Hope that doesn't screw up the performance of...

Here is the list of fully compatible functions (any signature that could varmat, does varmat, 320 functions): [fully_compatible.txt](https://github.com/stan-dev/cmdstan/files/5965917/fully_compatible.txt) Here is the list of functions which have at least one signature...

@dashadower @hyunjimoon give this a go and tell me if these instructions are clear or not. I'm not super happy with the formatting on the font size/formatting on the screenshots....