Model,BERTScore-F1,BLEURT,Perplexity,color_type "Base do_sample=False",-0.087,0.4176,1065.0,secondary "Instruct do_sample=False",-0.087,0.4176,612.0,secondary "Instruct do_sample=True",0.17051547928713262,0.4681127144023776,612.2593940666384,baseline Fine-tuned,0.24532897425815464,0.4870651926845312,370.53503376170977,primary