1 paper · 1 filter
Eric Zhang, Leshem Choshen, Jacob Andreas
There are two primary ways of incorporating new information into a language model (LM): changing its prompt or changing its parameters, e.g. via fine-tuning. Parameter updates incu…