Diffrent ICC results between summ commend and sjPlot

[Daph2678]

I am running a MLM model using lmer with an additional level 1 variable.
The ICC I get using summ is very different than what I get using sjPlot (and also regular flowup ICC request after summary).

Why is that? Perhaps it takes into account not just the groups' random effect but also the added variable somehow?

I prefer to use the jtool summ results cause they are better for my aim, but not sure I can relay on it is because of the difference.