NeuPL is an efficient general framework that learns and represents policies in symmetric zero-sum games within a single conditional network.