Our work on Autoencoding Probabilistic Circuits has been accepted to the Eighth Workshop on Tractable Probabilistic Modeling (TPM) 2025! We introduce autoencoding architectures for probabilistic circuits enabling unsupervised representation learning.