The gemara 61b made a distinction between רבית קצוצה and אבק רבית. The beis din forces the return of ribbis ketzutza but does not force the return of avak ribbis (tosafos 62a cites 2 opinions if there is any point in returning ribbis ketzutza according to the opinion that it need not be returned. Perhaps avak ribbis which according to everyone does not NEED to be returned, would also be dependent on those 2 opinions whether it is proper to return. There is also a machlokes rishonim whether the borrower can grab back avak ribbis or not - see rosh).
Ribbis Ketzutza would be when the ribbis is an agreement in the loan, lending $10 and obligating the borrower to return $15. Avak Ribbis is any situation where the interest isn't guaranteed. Rashi 62b explains that when the lender takes a field as collateral and eats the fruits without deducting a penny from the loan, it is only avak ribbis since there is a possibility of the land not producing any fruits that year so that the ribbis isn't guaranteed. Included in avak ribbis would be all cases of buying and selling, when there isn't an actual loan.
Aside from these 2 categories of Ribbis, the gemara also introduced a category called הערמת רבית, which the gemara 62b explained is a lower level than avak ribbis and is only prohibited because it "looks like ribbis". This only applies to a situation where there really isn't any ribbis at all, because a technical trick was done to avoid the ribbis.
The gemara on 64b speaks about the borrower renting to the lender at a discount as an appreciation for the loan. Although the mishna says that this is assur, rashi doesn't explain the problem to be avak ribbis, rather he explains it to be מחזי כרבית. It seems that rashi understands that avak ribbis can only apply when the borrower it losing something by paying the lender, but here the house isn't standing to be rented out anyway so the borrower isn't losing anything - therefore, it only looks like ribbis.
The gemara also has a category of צד אחד ברבית which means that it is a chance of developing into a ribbis problem. This applies to any situation where the borrower or lender have the ability to avoid the ribbis problem (tosafos 63a d.h. tzad). However, this is only the opinion of R. Yehuda, the chachamim maintain that this is an issur of ribbis (and can even be d'oraysa - as tosafos 63a d.h. ribbis, indicates).
In our gemara, when the borrower gives his fiels as collateral to the lender and makes a deal that if he doesn't pay back within 3 years the lender can keep the field - all the fruits that the lender ate would be ribbis. Rashi introduces a new category called מחזי כרבית קצוצה. Normally something which just looks like ribbis doesn't need to be returned because it is a lower level than even avak ribbis, but here rashi says that since it "looks like ribbis ketzutza", it must be returned. The problem with rashi here is that rashi says that avak ribbis which doesn't need to be returned only applies to cases of buying and selling, but doesn't apply to cases of loans. Rashi seems to imply that all cases that begin with a loan are categorized as ribbis ketzuta or looking like ribbis ketzutza and needs to be returned. This contradicts the rashi on 62b who says that whenever their may not be ribbis (because the field may not produce) it would not be considered ribbis ketzutza.