A finetune ontop of the original Francois-PE model that incorporates KTO to increase coherency and prose. The model aims to have short and sweet prose