Safety · Ars Technica ·
These LLMs are the best at resisting Russian propaganda
An Estonian government benchmark compares dozens of large language models on their ability to resist Russian propaganda and strategic narratives. The results identify which models were most effective at avoiding misleading content.