While this seems to be a bad idea, I can see two ways of doing it :
1. Searching for < followed by anything but the % character, then ignoring it
(<)(?:[^%])
The [^] sequence allows you to search for anything but the following character.
The (?:) sequence is for non capturing groups.
2. (Better, if supported) Searching for input not followed by % with a negative lookahead
<(?!%)
The (?!) sequence succeeds if it doesn't match the following character, but is not captured.
If you also want to do it for %>, you can just "reverse" the first option :
(?:[^%])(>)
Or you need a negative lookbehind :
(careful here, the lookahead won't work as you need to go backwards)
(?<!%)>
与恶龙缠斗过久,自身亦成为恶龙；凝视深渊过久,深渊将回以凝视…
