Alan JoAttention Output, 최종 값 (Context Vector)
Query, Keys, Value할 때 value와 다르고 거기에 attention weight 곱해서 나오는 게 attention value. 헷갈리니까 Attention output이라고 하자
Each word is also represented by a value which contains the information of that word. As a result, each context word is now represented by an attention-based attention value of all the words in the sentence.
