SpaRC and SpaRP: Spatial Reasoning Characterization and Path Generation for Understanding Spatial Reasoning Capability of Large Language Models

Anonymous

Abstract: Spatial reasoning is a crucial component of both biological and artificial intelligence. In this work, we present a comprehensive study of the capability of current state-of-the-art large language models (LLMs) on spatial reasoning. To support our study, we created and contribute novel spatial characterization frameworks and datasets, the Spatial Reasoning Characterization (SpaRC), and Spatial Reasoning Paths (SpaRP), to enable an in-depth understanding of the spatial relations and compositions as well as the usefulness of spatial reasoning chains. We found that all state-of-the-art LLMs do not perform well on the datasets---their performances are consistently low across different setups. The spatial reasoning is an emergent capability as model sizes scale up. Finetuning both large language models (e.g., Llama-2-70B) and smaller ones (e.g., Llama-2-13B) can significantly improve their F1-scores by 7--32 absolute points. We also found that the top proprietary LLMs still significantly outperform their open-source counterparts in topological spatial understanding and reasoning.

Paper Type: long

Research Area: NLP Applications

Contribution Types: Model analysis & interpretability, Data resources, Data analysis

Languages Studied: English