I've figured out that the problem is the lower numerical precision of mexp() as compared...
This bug still seems to exist in JAGS 3.4. I'm running a similar Gaussian process...