Imagine I have a table like the following one.

set.seed(12)
table = 
  data.frame(
    value = rnorm(n = 10), 
    par = runif(n = 10, min = - 1, max = 1)
  )

How can I extract the entries of value and par that correspond to the two smallest values of par above zero and the two biggest ones below zero? What I would like to obtain is something like

out = 
  data.frame(
    value = c(-0.2722960, -0.1064639, -0.3153487, 0.4280148),
    par = c(-0.464112814, - 0.121141350, 0.009535904, 0.339638592)
  )

I would appreciate if this could be done using dplyr so to be able to do it for bigger dataframes with grouping variables.