Why can't I pass an anonymous function to dplyr's filter without getting errors about vector size?

Question

Coming from base R, I expected the following to work as a way to pass anonymous functions

library(tidyverse)
starwars %>%
  select(height) %>% 
  filter(function(x) x > 100)

It does not. It reports the following:

Error in `filter()`:
! Problem while computing `..1 = function(x) x > 100`.
Caused by error in `vec_size()`:
! `x` must be a vector, not a function.
Run `rlang::last_error()` to see where the error occurred.

I suspect that I'm making a fundamental error, but I see no comparable examples in the documentation. Where is my misunderstanding of how dplyr handles such cases?