# Improving Weak World Models Behind Strong Agents in Atari Pong

arXiv: https://arxiv.org/abs/2607.15142 · HTML: https://mlsanity.com/paper/2607.15142/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Latent dynamics models and model-based RL, one of 11 directions in World models.
- Builds on 5 of its 21 references that are in the map.
- First posted on arXiv 16 Jul 2026.

## Placed in

- World models / [Latent dynamics models and model-based RL](https://mlsanity.com/wm/latent-dynamics-models-and-model-based-rl/)

## Builds on (most cited first, 5 of 5)

- [Proximal Policy Optimization Algorithms](https://mlsanity.com/paper/1707.06347/) (2017)
- [SAM 2: Segment Anything in Images and Videos](https://mlsanity.com/paper/2408.00714/) (2024)
- [Mastering Diverse Domains through World Models](https://mlsanity.com/paper/2301.04104/) (2023)
- [Retentive Network: A Successor to Transformer for Large Language Models](https://mlsanity.com/paper/2307.08621/) (2023)
- [Simulus: Combining Improvements in Sample-Efficient World Model Agents](https://mlsanity.com/paper/2502.11537/) (2025)

## Built on by (most cited first, 1 of 1)

- [RolloutFaith: Auditing Persistent Internal Interventions in Visual World Model](https://mlsanity.com/paper/2609.36843/) (2026)

## Abstract (arXiv)

Strong world-model agents frequently contain weak world models. We study this agent-world-model gap by reproducing five visual world-model agents in Atari Pong: DreamerV3, DIAMOND, TWISTER, Simulus, and STORM, with performance comparable to the reported results, and independently evaluating their frozen world models. First, closed-loop rollout diagnosis qualitatively inspects visual trajectories generated by each frozen model under an independently trained policy. All five models exhibit clear visual or dynamical failures, including ball disappearance, incorrect motion, and invalid ball-paddle interactions. Second, under native zero-shot model-based reinforcement learning (MBRL), a new policy is trained entirely within the frozen model from scratch using the agent's native RL procedure, without real-environment training. When evaluated in the real environment, these policies substantially underperform the reproduced agents: DreamerV3 (-5.5 to -20.9), DIAMOND (19.7 to -9.6), TWISTER (17.7 to -13.3), Simulus (20.8 to -11.6), and STORM (18.7 to -21.0), where -21 is the minimum Pong return. This gap also extends broadly across Atari100K. Motivated by the ball-related rollout failures in Pong, we propose Concept-Guided Spatial Regularization (CGSReg), an auxiliary reconstruction loss on task-critical concept regions. We evaluate it under a more challenging pixel-space zero-shot MBRL setting, where policies learn directly from images generated by the frozen world model. Ball-region CGSReg improves pixel-space zero-shot MBRL in DreamerV3, DIAMOND, TWISTER, and Simulus, and also improves closed-loop rollouts in the first three; STORM shows no clear improvement.