I'm not sure what you're asking. If you're working with n-bit integers and you interpret the value as unsigned, then all values >= 2^(n-1) (i.e. the upper half

I'm not sure what you're asking. If you're working with n-bit integers and you interpret the value as unsigned, then all values >= 2^(n-1) (i.e. the upper half of the range) represent negative two's complement values. Or put another way, bit n-1 is on iff the value is negative.