out-of-manifold-learning-strats Studying the effects of different replay strategies and feedback schemes on an RNN's "neural" manifold when adapting to changing tasks