&
Pages using &# in wikicode are sometimes changed to use actual unicode. This is usually fine, but in some cases this is needed to prevent normalization.
