Ultimately, I did some reading and found out that I need to use the conditions argument in the conditional_effects() function and set the grouping factor to NA. This post helped me get an answer.
I did this :
test <- conditional_effects(
modA2, method = "fitted",
robust = FALSE, re_formula = NULL,
effects = "Zcumul_xp",
conditions = data.frame(predator_id = NA)
)
And now I think I have the proper trend :
Maxime