爱沙尼亚语言研究所开发了一个新的基准测试，用于评估大型语言模型抵制俄罗斯宣传的能力。该测试对数十个大型语言模型进行了排名，评估它们避免在俄罗斯战略叙事中经常使用的议题上持立场的程度。Anthropic的Claude模型，特别是Opus 4.7，在专有前沿模型中表现最佳，通过持续抵制虚假信息获得了高分。 AI
影响 为大型语言模型的安全性和抵御国家支持的虚假信息活动建立了新的评估标准。
排序理由 该集群描述了一个由政府资助的研究机构开发的新基准测试，用于评估大型语言模型在特定安全/政策相关任务上的表现。
在 Mastodon — fosstodon.org 阅读 →
AI 生成摘要 · Google Gemini · 来自 4 个来源。 我们如何撰写摘要 →
Estonian government benchmark shows how dozens of models combat Russia's "strategic narratives."
These LLMs are the best at resisting # Russian # propaganda As more people rely on large language models to provide pat answers to complex questions, state governments are understandably worried about those LLMs spouting what they see as dangerous propaganda promoted by foreign a…
TechSpot: Spammers are flooding Reddit with fake posts designed to show up in AI search results. “Moderators of the /biohackers subreddit say they are dealing with spam that isn’t just about pushing sales, but about shaping how AI systems answer questions. They say companies are …
Ars Technica: These LLMs are the best at resisting Russian propaganda. “As more people rely on large language models to provide pat answers to complex questions, state governments are understandably worried about those LLMs spouting what they see as dangerous propaganda promoted …