Evaluating Language World Models in Structured Multi-Agent Sandbox Worlds.
Code | Paper | Dataset: In preparation