The forked copy of the respiratory was appending the first index of the most recent prediction. This is fine if and only if the truncated backpropagation length is one since there will only be one prediction; however, if the backpropagation length is greater than one this would result in the wrong prediction being appended to the prediction list. Therefore, my solution gets the most recent prediction; thus, removing the bug.
The forked copy of the respiratory was appending the first index of the most recent prediction. This is fine if and only if the truncated backpropagation length is one since there will only be one prediction; however, if the backpropagation length is greater than one this would result in the wrong prediction being appended to the prediction list. Therefore, my solution gets the most recent prediction; thus, removing the bug.