Attention based seq$-2-$seq model

Which of the following statements about the attention$-$based sequence to sequence model is correct? $($Note: More than one option may be correct.$)$

The most important output state of the encoder is calculated and passed as the context vector to the decoder.

All the output states of the encoder are passed and made available while decoding.

The value of the attention score can be calculated based on multiple methods.

All of the above