Hello, is there a known bug with the "text_match_score", I would expect that the second product here would be listed first, given that it matches brand and matches more tokens on "name": Reproduction: https://github.com/reinoldus/typesense-reproduction-repo/tree/token-count-weird (data in thread)