(i know this is technically v3, sue me)
This time I added all of the score tags, so outputs should be generally higher quality.
Prompt format is "score_9, score_8_up, score_7_up, score_6_up, score_5_up, score_4_up, 2000s emo, (tags)"
Dataset was the same as the previous model, but with all the score tags added & using a different autotagging method.