Smallworlds: Assessing Dynamics Understanding Of World Models In Isolated Environments

Abstract

Current world models lack a unified and controlled setting for systematic evaluation, making it difficult to assess whether they truly capture the underlying rules that govern environment dynamics. In this work, we address this open challenge by introducing the SmallWorld Benchmark, a testbed designed to assess world model capability under isolated and precisely controlled dynamics without relying on handcrafted reward signals. Using this benchmark, we conduct comprehensive experiments in the fully observable state space on representative architectures including Recurrent State Space Model, Transformer, Diffusion model, and Neural ODE, examining their behavior across six distinct domains. The experimental results reveal how effectively these models capture environment structure and how their predictions deteriorate over extended rollouts, highlighting both the strengths and limitations of current modeling paradigms and offering insights into future improvement directions in representat

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keyli2025smallworlds

Related papers

• Benchmarking World-model Learning (2025)1.57
• Dynamic Sparsity: Challenging Common Sparsity Assumptions For Learning World Models In Robotic Reinforcement Learning Benchmarks (2025)0.00
• Towards Unraveling And Improving Generalization In World Models (2024)0.00
• Foundation World Models For Agents That Learn, Verify, And Adapt Reliably Beyond Static Environments (2026)0.00
• World Models As An Intermediary Between Agents And The Real World (2026)0.00
• STORM: Efficient Stochastic Transformer Based World Models For Reinforcement Learning (2023)4.52
• The Effectiveness Of World Models For Continual Reinforcement Learning (2022)0.00
• Smaller World Models For Reinforcement Learning (2020)0.00