Awesome World Model Hub Papers · Datasets · Projects
← Back to datasets

AGI Maze as a Benchmark Framework for World-Modeling Agents

2026 Datasets
Preview for AGI Maze as a Benchmark Framework for World-Modeling Agents

Dataset Analysis

Introduces AGI Maze, a benchmark for evaluating world-modeling in LLMs via grid-based mazes requiring memory and hidden state representation.

Provenance

Collected from papers.

Derived from paper: AGI Maze as a Benchmark Framework for World-Modeling Agents

Resources