What is the difference between radix trees and Patricia tries?

Given the talk on compressing Wordle data. I thought people might be interesting an older method based on Patricia tries.