Umm transcribing 20 and 09 to 0 and 1 leads to 2 different results if you include or not the spaces between the text.
I think that it should be included as doing so gives you a result starting with "0a 54" or "\nT".
But in any case there are too much white spaces, so probably we need to find a way or removing some of them.
If the result is ciphered with rabbit then the output should be more random so it is almost impossible to get so many 0.
If the result is plaint text it can have some bias but still it is too high (almost 4 to 1).
Also you can try inverting 20 and 09 values but the it seems that 20 as 0 gives a more text like result.
Transcription without the spaces between words (hand made)
10100101 01000000 10000000 01000010 00000010 00000010 00100001 00010101 10010000 10000010 10000101 01001000 00001000 00100000 10001000 10010010 00000010 00000100 00101000 00001000 00000000 00010001 10000100 00001000 00010000 00010100 10010000 01000000 10001000 10011000 00001000 00001000 01000000 10100010 00000010 00010001 01000001 00000001 00001001 00000010 00000100 01000000 10000001 01100010 00000010 00100001 00001100 00000010 11000010 00000100 11000100 01000000 00100000 01000010 00000010 00000010 00001000 00010100 00001000 00001000 01100000 00000000 10000001 00000100 00000100 00001000 00001000 10000000 00100100 00011000 01000001 00000001 00010100 00001001 00010000 00010001 11000000 01001011 00000010 00000100 00000100 00001000 10000100 00000101 00000000 00000100 00010000 00010110 00001100 00100000 00110000 10000100 10000100 00100000 01000000 01000000
Transcription with the spaces between words (automated)
00001010 01010100 00001000 00000000 01000010 00000010 00000010 00100001 00000010 11001000 01000001 01000010 10100100 00000100 00010000 01000100 01001000 00000000 00100000 01000010 10000000 10000000 00000000 00100011 00001000 00010000 00100000 00000100 10010000 01000000 10001000 10011000 00001000 00001000 01000000 10100010 00000000 00100001 00010100 00010000 00010000 00010010 00000100 00001000 10000001 00000000 01100010 00000010 00100001 00001100 00000010 11000010 00000100 11000100 01000000 00000010 00000100 00100000 00100000 00100000 00010000 00101000 00010000 00010000 11000000 00000000 01000000 10000010 00000010 00000100 00000100 01000000 00010010 00001100 00100000 10000000 10001010 00000000 00100100 01000000 01000110 00001000 00001001 01100000 01000000 00100000 00100000 01000100 00100000 00101000 00000000 00100000 10000000 10110000 01100001 00000000 10000100 00100100 00100001 00000010 00000010 00010000
The text ends with 09 and 3 null characters (000000) and in both cases the length is too short, event if it is ciphered that's imposible as it works in blocks of 16 bytes, that's another reason that leads me to think that we should remove some bits.