>>106479383
Good afternoon /aicg/

Did a test of my personal 8B model fine-tune with a RP dataset, completing off of this prompt: files.catbox.moe/yeh1n0.txt

Obviously not perfect. Obvious logical fuckups, but noticeably better and imo not too bad for a 3B quanted finetune. How would you rate this one? Read the TXT file in order for the response to make sense.