GitHub

https://github.com

Updated 11 months ago

geobenchx • Science 54%

LLM-agents benchmark set of geospatial tasks requiring multi-step tool use; and LLM-as-Judge based evaluation framework.