AGI Maze Prediction Datasets: A Compact Benchmark for Learning World Dynamics with Transformers
ResearchAI 60
Source: arXiv cs.AIPublish time unverified
arXiv:2609.02339v1 Announce Type: cross Abstract: World modeling requires a predictive model to maintain and update an internal state adequate for reasoning about the consequences of actions. We introduce the AGI Maze Prediction Datasets and Benchmark, a lightweight controlled testbed for studying this capability in Transformers and other predictive models.