It is in fact a set difference and the sets in question are values for the column being ONed on in each table (as I explained in the top level comment)
My reply was with respect to these being cross joins with filters. That’s just one way of thinking about a JOIN that actually adds notational complexity (and does not actually represent how your database would be doing the join because it would be unnecessarily expensive computationally). Thinking of it in terms of a set difference of the values in thr ON column is arguably the better way of thinking of it if you are interested in how such a join should be implemented. Alternatively you can interpret the venn diagrams as just the filtration if you prefer interpreting as a cross join with filters.