This will have the xgen-7b-8k-inst LLM

xgen lacks context in what the user said in prompt, that can be fine tune or changed a bit in the training data