Equivalent of R-squared in negative binomial regression

Question

In my study, I experiment with fixed and mixed effects negative binomial regression to my data (in R) as the response variable is a count variable. I have read somewhere that unlike in the case of linear or logistic regression, we can not rely on R-squared in negative binomial regression to assess the proportion of variance explained by the model. Instead, it is suggested that we should use McFadden's pseudo R-squared or maximum likelihood R-squared, but their interpretation is not as straightforward as R-squared. However, in some studies I noticed using marginal or conditional R-squared with mixed effects negative binomial regression. My question: what is an optimum metric then and how to interpret it?