World models enable language‑model agents to predict environment dynamics and plan before acting. In text environments the model must learn symbolic action effects from serialized state descriptions, yet the impact of serialization structure has been underexplored. We introduce HyperWorld, a controlled study of state serialization for learned textual world models. We compare raw observations, independent sentences, pairwise triples, and entity‑centered hyperedge units that group multiple related facts around entities and relations. All variants share the same training objective: given a state and an action, predict symbolic effects or judge the action infeasible.
Across model scales, data budgets, and both in‑distribution and out‑of‑distribution test worlds, hyperedge serialization yields the clearest gains for 0.5B‑1.5B models and under distribution shift. Larger models narrow the gap, and pairwise triples can match or slightly exceed hyperedges on in‑distribution exact match, but hyperedges achieve the strongest out‑of‑distribution fact F1 and the best trade‑off between feasibility detection and effect prediction for small‑to‑medium models. In downstream greedy planning, the hyperedge world model also attains the highest success rate among the tested representations.
These findings demonstrate that higher‑order state organization is a simple yet effective inductive bias for learned symbolic world models, especially when model capacity is limited or test environments differ from training.
Review: Hypergraph‑based serialization supplies rich structural cues that keep textual world models robust under resource constraints and distribution shifts, making it a promising design choice for future language‑agent systems.