Both take a theta and an x whose rows are independent observations,
check them, standardize them, and either return the n_theta x n_obs matrix
or its row sums. Which functions do the scoring, and whether the Jacobian
comes with them, is surrogate_ops()'s business.
Arguments
- fit
An
nsbi_nleornsbi_nrefit.- theta, x, sum_iid, max_batch
As in
log_lik().