This is the question:

By definition of the IEEE754 standard, 32-bit floating point numbers are represented as follows:

S (1 bit) E (8 bits) M (23 bits) S: Sign bit E: Exponent M: Mantissa

Which of the following is the correct “mask bits” in hexadecimal to be used for extracting only the exponent part of the above format?
Here, “mask bits” means a bit pattern which is logically ANDed withthe 32-bit floating point value.

a) 107FFFFF b) 7F800000 c) FF100000 d) FF800000

The given answer is b : 7F800000, but i have no idea why, can anyone give me an explanation, i'm greatly appreciated!

Is that because :

7F800000 = 0111 1111 1000 (8)

While others answer contain more "1" in binary format?

When you bitwise AND a number in this format with 7F800000 you get:

    SEEEEEEEEMMMMMMMMMMMMMMMMMMMMMMM
AND 01111111100000000000000000000000

= 0EEEEEEEE00000000000000000000000


Everything gets set to 0 except the exponent. Yes, it is necessary for the correct answer to have 8 ones, but it is also necessary that they are in the correct positions. 00007F80 would not be correct because it would result in:

    SEEEEEEEEMMMMMMMMMMMMMMMMMMMMMMM
AND 00000000000000000111111110000000

= 00000000000000000MMMMMMMM0000000