The most popular way we evaluate large language models measures the wrong thing: likeability over accuracy and value.
In pet genetics, cancer research, and beyond, Charlie Lieu, MBA ’05, SM ’05, has spent her career harnessing massive data ...