Neural quantum states (NQS) are a promising approach to study many-bodyquantum physics. However, they face a major challenge when applied to latticemodels: Convolutional networks struggle to converge to ground states with anontrivial sign structure. We tackle this problem by proposing a neural networkarchitecture with a simple, explicit, and interpretable phase ansatz, which canrobustly represent such states and achieve state-of-the-art variationalenergies for both conventional and frustrated antiferromagnets. In the lattercase, our approach uncovers low-energy states that exhibit the Marshall signrule and are therefore inconsistent with the expected ground state. Such statesare the likely cause of the obstruction for NQS-based variational Monte Carloto access the true ground states of these systems. We discuss the implicationsof this observation and suggest potential strategies to overcome the problem.