-nan results with lkj_corr_cholesky

[bgoodri]

For Cholesky factors that big, it is possible that the diagonal elements result in the square root of a slightly negative number when symbolically they should be slightly positive. If you search for the “onion method” on the Stan-users Google group, there is some code for a more numerically stable way to construct a Cholesky factor of a large correlation matrix.