Text this: Model mimicry limits conclusions about neural tuning and can mistakenly imply unlikely priors