Deep generative models have achieved remarkable success in generating diverse and coherent textual content. Recently, there has been growing interest in exploring the potential of binary representations for encoding and decoding text. This approach leverages the inherent efficiency and computatio