fix markdown links to various image files that use a %%TOKEN%%

2013/cable2/index.html

@@ -469,7 +469,7 @@ this entry win the IOCCC, it will be the first time (to the author’s knowledge
 that the IOCCC has been used for peer review/publication of original research.</li>
 <li>OCR in general is a hard problem. Here, we only recognise one possible form
 for each character. You can see what the supported character shapes look like by
-looking at the examples in <a href="https://github.com/ioccc-src/temp-test-ioccc/blob/master/2013/cable2/ascii.bmp">ascii.bmp</a>. If you copy the shape of the
+looking at the examples in <a href="ascii.bmp">ascii.bmp</a>. If you copy the shape of the
 character forms you find in there, recognition accuracy should be very good
 (95%+). A more “useful” version would extend the character stroke table to
 support multiple different commonly-used forms for each character.</li>