The short answer is that the EPCs are reported in hexadecimal (also called hex) format by default.
A 96-bit EPC is a long binary number, for example: 001100000010100000110010110000100001100001001100000000100010100100111001100100001111010000011101.
For convenience, this long binary number is usually represented in hexadecimal (base 16, or hex) form. In hex format the digits {0,1 2, 3, 4, 5, 6, 7, 8, 9} in hex format are represented by the familiar {0, 1 2, 3, 4, 5, 6, 7, 8, 9} characters, while the values {10, 11, 12, 13, 14, 15} are represented by the characters {A, B, C, D, E, F} or (a, b, c, d, e, f} respectively. To represent a binary number in hex format 4 bits are be grouped together to define a single hex digit. Four binary bits can have 16 values: 0000, 0001, 0010 ... 1110, 1111. These values are represented in hexadecimal as 0 to F. So 0000 is represented by 0, 1001 is represented by 9 and 1101 is represented by D.
Thus the long binary number above becomes 302832C2184C02293990F41D in hex. Finally, to aid readability a hyphen is often added between every 2 characters. So, all of these represent exactly the same EPC:
1. 001100000010100000110010110000100001100001001100000000100010100100111001100100001111010000011101,
2. 302832C2184C02293990F41D,
3. 30-28-32-C2-18-4C-02-29-39-90-F4-1D.