Unlearning Climate Misinformation in Large Language Models

TL;DR

This paper explores methods to improve the factual accuracy of large language models on climate change topics, focusing on unlearning misinformation and evaluating their robustness against false data.

Contribution

It introduces effective unlearning algorithms for climate misinformation in LLMs and compares their performance with fine-tuning and RAG techniques.

Findings

Unlearning algorithms can effectively address nuanced climate claims.

Poisoning with false information may not impact responses in other domains.

Unlearning shows promise despite previous limitations in privacy contexts.

Abstract

Misinformation regarding climate change is a key roadblock in addressing one of the most serious threats to humanity. This paper investigates factual accuracy in large language models (LLMs) regarding climate information. Using true/false labeled Q&A data for fine-tuning and evaluating LLMs on climate-related claims, we compare open-source models, assessing their ability to generate truthful responses to climate change questions. We investigate the detectability of models intentionally poisoned with false climate information, finding that such poisoning may not affect the accuracy of a model's responses in other domains. Furthermore, we compare the effectiveness of unlearning algorithms, fine-tuning, and Retrieval-Augmented Generation (RAG) for factually grounding LLMs on climate change topics. Our evaluation reveals that unlearning algorithms can be effective for nuanced conceptual claims, despite previous findings suggesting their inefficacy in privacy contexts. These insights aim to guide the development of more factually reliable LLMs and highlight the need for additional work to secure LLMs against misinformation attacks.