Upon searching for a term, a number of tags are presented (each with given a different weight which represented by fontsize) along with search result (e.g. the following is the search result for search term 'Jesus'
http://www4.wiwiss.fu-berlin.de:8800/search/search.aspx?QueryString=Jesus&am...)
It's interesting that, for the sample search on "Jesus" as shown above, one of the biggest-print tags is "Fictional character".
There seem to be character-set problems; Wikipedia is served under UTF-8, and so is DBpedia, but something in between seems to be mangling the non-ASCII characters in search results so you end up with random garbage.