GREAT genome ontology top hits criteria?

Question

[a_bis]

Hi, when running a GREAT analysis of a bed file using the default settings, what are the criteria by which the top few hits are shown in each of the gene ontology categories? For example, in 'mouse phenotype single KO' I get 8 terms on the results page, but then when I download 'all data' in .tsv format, I get a file several hundreds of lines long. In fact, in some of these lines the 'BinomP' value and the number of genes associated with a particular GO term are lower and higher respectively than the few hits from the original result page. Any ideas what the criterion of display in the original results page is? Thanks very much!