Multivariate Laplace Prior

Thanks for the input! Then, for the logp, should I just implement the log of the likelihood function listed on the wiki? If so, how do I implement the Bessels function? Is there an option for that in Pytensor?