Currently the evaluator doesn't differentiate between int and float. You can resolve either from the same variable. There is ongoing discussion on type checking f.e.: open-feature/flagd#1481. After community reaches some agreement, we should verify if our implementation follows it.