Hello, yes we did run the find training data step a few times and labeled some records, sorry about the bad quality, but it shows 66 as positives and 38 as not matches.
Is there anywhere we can look to try and do some debugging about that might be happening? i can give you some context about data if thats useful