The machine epsilon for float.
It is the smallest possible and reliable number so that floatValue + Epsilon > floatValue holds true, whereby the result remains precise.
floatValue + Epsilon > floatValue
true