Censored likelihoods with different bounds per observation

Questions v5
2

The bounds can have the same shape as your observed data just fine, they don’t need to be scalar.