Most LLMs cannot reliably evaluate text on the level of individual letters. A technique called byteification retrofits ...
10d
When I added a newline to regexes checking for 'digits only', 9 out of 14 implementations passed it
I provided 14 different ways of 'checking if a string is digits only' with a string consisting of 123 followed by a newline.
Some results have been hidden because they may be inaccessible to you
Show inaccessible results