@tszzl @repligate that alone would explain a lot? there’s a high chance that rules in datasets are at least contradictory, if they are hard to curate / bring to a unified vision that could be communicated to the model
in reply to: 1994990907301527937
same thread: 1994982195359092784 1994984560883335666 1994987754523169272 1994990198816694629 1994990907301527937 1994991487134962046 1994993173815595091 1994994711673921915 1994995072396660903 1995001025200296128 1995020049074323613 1995020142808547502 1995023774207676501 1995325209357009277
Reproduced against link rot, credited and linked to its original. Yours and you’d rather it weren’t here? Open an issue.