RISE Research Radar

Computer Science Open House 2022-2025

2025

ClimateEval: A Comprehensive Benchmark for NLP Tasks Related to Climate Change

Murathan Kurfali, Sherong Zuhre, Johan Neiver, Gabriela Hascova

Summary

Presents unified benchmark aggregating climate-related NLP evaluation datasets. Introduces Guardian Climate News Corpus with 6 climate-related categories. Includes 13 datasets covering stance detection, claim verification, news classification. Evaluates open-access baseline LLMs (2B-70B parameters) finding models struggle with domain-specific entities and seemingly simple tasks.

Themes

llmsustainability

Keywords

NLP benchmark, climate change, text classification, claim verification, LLM evaluation

Poster

ClimateEval: A Comprehensive Benchmark for NLP Tasks Related to Climate Change poster

Click image to open full size