Clarify how "valid floating-point number" is used

Add a description of Clarify how "valid floating-point number" is used