From the document said that the option --count going to
prefix lines by the number of occurrences
but after I do some test it seems doesn't work.
$ echo -e "a\na\b\nc" |uniq --count
1 a
1 a
1 c
Should the line 'a' occurs 2 times?
If I understand what you are trying to do, there is a typo in your original string that is confusing uniq (it thinks there are 1+1+1 lines instead of 2+1+1). There is an "n" missing.
I guess this is the result you were expecting, as it is coherent with the man page.