Trasformers, how does positional encoding apply in n dimension enconding of a word?

Question

Imagine we have the first word of a sentence rapresented with a vector such as [1,2,3,4], and the second word with a vector such as [5,1,6,4]. The questions is: if we apply