Rnn_cell_backward

First, you posted this in Convolutional Neural Networks (DLS course 4) but your query belongs to DLS course 5. Kindly change the category as instructed here.

So, what is sq?

Hint: In a notebook, it is said that “\tanh(W_{ax}x^{\langle t \rangle}+W_{aa} a^{\langle t-1 \rangle} + b_{a}) was computed and saved as a_next in the forward pass.”

Also, note that sharing your code is not allowed. Just sharing your full error will suffice.