I'm looking for games from famous grandmasters, in which pawn promotion to a rook, a bishop, or a knight is a better move than queen promotion. I tried to Google some, but I couldn't find anything.
Knight promotion is reasonably common (though not nearly as much so as Queening); Rook and especially Bishop, much less so, because the only motivation is stalemate avoidance(*). Tim Krabbé has a collection of underpromotion in serious play. You can judge for yourself how famous the players are in each case: https://timkr.home.xs4all.nl/chess2/minor.htm
(*) In principle one could also underpromote in order to get stalemated. This does not seem to have happened yet in actual play, but the same Krabbé page shows several examples of composed studies where such an underpromotion is the only drawing move.
P.S. Further Googling brings up a page that cites a drawing N-promotion in "a game between Michael Adams and Tony Miles at Tilburg on 17 November 1993": https://www.chesshistory.com/winter/extra/underpromotion.html
These players probably qualify as sufficiently famous . . .
Here is a list of underpromotions in famous games: https://www.chess.com/article/view/9-amazing-underpromotions
Underpromotion to a knight is not that rare, especially in cases where the new knight gives a check or fork. But bishop / rook underpromotions are exotic, usually to avoid a statemate.
This is the "trap line" of Albin Counter-Gambit. at 3:10, you can see that promoting to a Knight is better than promoting to a Queen, simply because promoting to Queen is too slow (not to mention it would be lost anyway), and Knight promotion delivers a check which lets you keep the initiative. It's a line that requires a lot of analysis, so if you are rated lower than 1600, you may feel overwhelmed by all the possibilities in this variation.