Can Factual Opinions Be Edited (Manipulated) in Large Language Models?

2026-06-01

AI Summary

This work addresses a critical gap in current knowledge editing techniques for large language models, which predominantly focus on atomic facts while overlooking the risks associated with editing factual opinions—such as public figures’ stances on societal issues. The study presents the first systematic investigation into the editability of such opinions, introducing the FOE benchmark comprising 261 public figures, 19 issue categories, and 2,178 opinion records. It further proposes a Self-Generated Evidence-Aligned editing method that, without explicit instructions, ensures consistency between edited opinions and the model’s self-generated supporting evidence. Experimental results demonstrate that existing approaches merely achieve superficial modifications, whereas the proposed method significantly enhances factual consistency, thereby establishing a new foundation for safe and controllable knowledge editing in large language models.

Abstract

Large Language Models (LLMs) are increasingly integrated into various domains, making knowledge editing techniques crucial yet potentially hazardous. Current editing methods primarily target atomic facts, overlooking the significant risks associated with manipulating factual opinions, e.g., documented stances of public figures on societal issues. Such manipulation could reshape public images, influence elections, and alter societal views. To systematically assess this threat, we introduce the Factual Opinion Editing with Evidence (FOE) benchmark, which encompasses 261 public figures, 19 issue categories, and 2,178 complete opinion records. Our evaluations demonstrate that current editing techniques struggle significantly with factual opinions, often achieving only superficial changes while failing to preserve consistency between the edited opinion and the supporting evidence generated by the model. To address this limitation, we further propose a simple yet effective Self-Generated Evidence-Aligned method that achieves opinion-evidence alignment without relying on explicit instructions. Together, our benchmark and method provide a foundation for understanding the emerging security implications of factual opinion editing in LLMs.

Problem

• Factual Opinion Editing
• Large Language Models
• Knowledge Manipulation
• Opinion Consistency
• AI Security

Innovation

• factual opinion editing
• knowledge editing
• evidence alignment
• large language models
• FOE benchmark