50 observations on each of 100 variables. The data are simulated such that
variables with similar indices are associated with the degree of association
decaying as variables are further apart (i.e., a correlation or information
matrix with have larger values near the diagonal). Details are given in the vignette.