PulseAugur
实时 13:41:16
English(EN) Microsoft claimed its new cybersecurity model scored 95.95% on CyberGym at half the cost, but the score didn't appear on the public leaderboard and independent

微软新的网络安全模型面临透明度质疑

微软推出了一款新的内部网络安全模型,声称其在 CyberGym 基准测试中取得了 95.95% 的分数,而成本仅为通常的一半。然而,该分数并未出现在公开排行榜上,且在发布前没有独立测试人员对其进行评估。此外,微软自己的文档显示,该模型在三个特定的漏洞利用类别中表现为无效。 AI

影响 引发了对关键安全应用中 AI 模型透明度和独立验证的质疑。

排序理由 该条目描述了一家主要科技公司发布新模型,但它不是前沿 AI 发布,而是侧重于特定应用(网络安全),而非通用 AI 能力。

在 Mastodon — sigmoid.social 阅读 →

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →

微软新的网络安全模型面临透明度质疑

报道来源 [1]

  1. Mastodon — sigmoid.social TIER_1 English(EN) · [email protected] ·

    Microsoft claimed its new cybersecurity model scored 95.95% on CyberGym at half the cost, but the score didn't appear on the public leaderboard and independent

    Microsoft claimed its new cybersecurity model scored 95.95% on CyberGym at half the cost, but the score didn't appear on the public leaderboard and independent testers never saw the model before launch. The company's own documentation shows zero performance on three exploit categ…