Text2World: Benchmarking Large Language Models for Symbolic World Model Generation

TL;DR

This paper introduces Text2World, a comprehensive benchmark for evaluating large language models' ability to generate symbolic world models from text, addressing previous evaluation challenges and providing insights into current capabilities and future improvements.

Contribution

The paper presents a new benchmark, Text2World, with diverse domains and robust metrics, and evaluates LLMs, revealing their limitations and exploring strategies to improve world modeling.

Findings

Reasoning models with reinforcement learning outperform others.

Even the best models show limited world modeling capabilities.

Strategies like test-time scaling and agent training can enhance performance.

Abstract

Recently, there has been growing interest in leveraging large language models (LLMs) to generate symbolic world models from textual descriptions. Although LLMs have been extensively explored in the context of world modeling, prior studies encountered several challenges, including evaluation randomness, dependence on indirect metrics, and a limited domain scope. To address these limitations, we introduce a novel benchmark, Text2World, based on planning domain definition language (PDDL), featuring hundreds of diverse domains and employing multi-criteria, execution-based metrics for a more robust evaluation. We benchmark current LLMs using Text2World and find that reasoning models trained with large-scale reinforcement learning outperform others. However, even the best-performing model still demonstrates limited capabilities in world modeling. Building on these insights, we examine several promising strategies to enhance the world modeling capabilities of LLMs, including test-time scaling, agent training, and more. We hope that Text2World can serve as a crucial resource, laying the groundwork for future research in leveraging LLMs as world models. The project page is available at https://text-to-world.github.io/.