EarthSE: A Benchmark for Evaluating Earth Scientific Exploration Capability of LLMs

TL;DR

This paper introduces EarthSE, a comprehensive benchmark with datasets and metrics to evaluate large language models' ability to perform scientific exploration in Earth sciences, covering fundamental to advanced tasks.

Contribution

It presents a new holistic benchmark with diverse datasets and evaluation metrics specifically designed for assessing LLMs' Earth science exploration capabilities.

Findings

Leading LLMs show significant limitations in Earth science exploration tasks.

The benchmark reveals substantial room for improvement in LLMs' scientific reasoning.

EarthSE provides a standardized platform for future research in scientific exploration evaluation.

Abstract

Advancements in Large Language Models (LLMs) drive interest in scientific applications, necessitating specialized benchmarks such as Earth science. Existing benchmarks either present a general science focus devoid of Earth science specificity or cover isolated subdomains, lacking holistic evaluation. Furthermore, current benchmarks typically neglect the assessment of LLMs' capabilities in open-ended scientific exploration. In this paper, we present a comprehensive and professional benchmark for the Earth sciences, designed to evaluate the capabilities of LLMs in scientific exploration within this domain, spanning from fundamental to advanced levels. Leveraging a corpus of 100,000 research papers, we first construct two Question Answering (QA) datasets: Earth-Iron, which offers extensive question coverage for broad assessment, and Earth-Silver, which features a higher level of difficulty to evaluate professional depth. These datasets encompass five Earth spheres, 114 disciplines, and 11 task categories, assessing foundational knowledge crucial for scientific exploration. Most notably, we introduce Earth-Gold with new metrics, a dataset comprising open-ended multi-turn dialogues specifically designed to evaluate the advanced capabilities of LLMs in scientific exploration, including methodology induction, limitation analysis, and concept proposal. Extensive experiments reveal limitations in 11 leading LLMs across different domains and tasks, highlighting considerable room for improvement in their scientific exploration capabilities. The benchmark is available on https://huggingface.co/ai-earth .