-4095  sign 1, exponent 5, significand 10

A. Treat as unsigned and convert to binary

 4095
 2047  1
 1023  1
  511  1
  255  1

  127  1
   63  1
   31  1
   15  1

    7  1
    3  1 
    1  1
    0  1    1111 1111 1111b

B. no .decimal portion

C. Adjust the binary value to scientific style notation.

   1.111 1111 1111 x 2^11

D. Calculate the bias for the exponent

  2^(5-1)-1 = 2^4-1 = 16-1 = 15

E. Add bias to exponent and covert to binary

   15 + 11  26

   26
   13  0
    6  1
    3  0
    1  1
    0  1   11010 

F.  
Sign negative so sign bit 1

Biased exponent      11010

Find significand

  Most significant bit 1. not stored

  And float can hold only 10 significant digits, so we store only the most 
  significant portion.

  1.111 1111 1111
  ^ ^^^ ^^^^ ^^^
  h  10 digits
  i
  d
  d
  e
  n

  1 11010 111 1111 1111