Similar to the percentage bug (changing numbers lead to 4 % > 5% %):
Sub-clause 16.2, delete the sub-clause
Ziffer 16.2: Ziffer streichen
Now there is a whole bunch of them, and as the phrase "delete the sub-clause" appears in several different contexts, it is difficult to sort it out with find and replace.
CT keeps proposing sentences with the same wording, but a different number as e.g.
Ziffer 13.1, Ziffer streichen
I assume that CT understands the comma as part of the number. But it shouldn't when there is a blank after it or when it is at the end of the segment. This leads – during the fuzzy processing – here and in many other use cases to wrong numbers with sudden commas and full stops where they do not belong.