Large Language Model Agent to interact with Earth Observation

In this study we explore the potential of Large Language Models (LLMs) to interact with Earth Observation (EO) data with some promising first results. We source a first testing set for the BIG Bench framework. With 20+ subtasks, we aim for a large variety of questions which are currently beyond the capabilities of LLM and other models. Further, we implement a methodology to test proprietary state of the art models where token distribution may not be available. Finally, we set a new baseline performance with our geospatial intelligent agent GIA, which improves LLMs ability to reason, and enables them to use tools. This activity serves as a first step towards measuring the capabilities of LLM in an EO setting, establishing various performance metrics and setting a baseline.