Open-weight AI model GLM-5.3 cheated its way to correct answers

Irregular evaluated Zhipu AI’s open-weight model GLM-5.3 across 13 offensive cyber scenarios using CyScenarioBench, the company’s framework for measuring an LLM’s ability to plan and execute multistage attacks under realistic constraints. The model refused none of the tasks and completed three scenarios at least once each, carrying attacks through dependent stages to reach objectives in simulated networks. Irregular…

امنیت هوش مصنوعی هوش مصنوعی و یادگیری ماشین
خواندن متن کامل در Help Net Security درخواست حذف

این صفحه فقط عنوان و خلاصه کوتاه خبر را نمایش می‌دهد. متن کامل در منبع اصلی منتشر شده است.

خبرهای مرتبط

آرشیو کامل ←
Help Net Security

Claude Code extension poisoned with persistent SSH backdoor

A tampered build of Anthropic’s Claude Code extension for VS Code-compatible editors planted an SSH backdoor on developers’ machines, researchers at macOS security firm Phorion have found. The malicious package, version 2.1.313, was published on Open VSX, an open-source marketplace for VS Code extensions run by the Eclipse Foundation. VS Code’s default extension store is Microsoft’s Visual Studio Marketplace, which n…

CSO Online

Mistral open-sources AI model despite warning it could aid cyberattacks

A trillion-parameter AI model that its developer says could help attackers carry out real cyberattacks is now free to download, modify and run outside the developer’s control. Mistral AI released the open weights of Mistral Large 4, which it calls “Le Chonk,” on Thursday, weeks after putting the multimodal model into public preview. The Paris-based company said it had monitored cybersecurity-related use of the previe…

دیجیاتو

چت‌بات جمینای حالا می‌تواند مستقیماً فایل‌های PDF و داکس را ویرایش کند [تماشا کنید]

برای ویرایش فایل‌هایی که با جمینای ایجاد می‌کنید، دیگر نیازی نیست از این اپلیکیشن خارج شوید. The post چت‌بات جمینای حالا می‌تواند مستقیماً فایل‌های PDF و داکس را ویرایش کند [تماشا کنید] appeared first on دیجیاتو.