LLMs compress U.S. climate opinion diversity and apply uniform gender assumptions that match reality for White and Hispanic Americans but misrepresent Black Americans.
Climate-Eval: A Comprehensive Benchmark for NLP Tasks Related to Climate Change, May 2025
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CY 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
How Large Language Models Systematically Misrepresent American Climate Opinions
LLMs compress U.S. climate opinion diversity and apply uniform gender assumptions that match reality for White and Hispanic Americans but misrepresent Black Americans.