Comment by hoansdz

What happens if the model continuously hallucinates or provides wrong answers until the user give up ? Will the AI trust those wrong answers and save it ?