Delta-Vector/Control-8B
Updated
•
108
•
2
finetunes to try and remove narration-slop from models and try to make RPs feel more "chatlike"
Note Experimental finetune to try and remove narration-slop from models and try to make RPs feel more "chatlike"?
Note DPO lora applied ontop to increase coherency
Note KTO lora applied ontop of Control-V1.1 - Overall improves prose and writing.