论文多源确认
Simon Willison 实测 GPT-6 Astra 直接解读模型权重的方法
There's a buzzy new paper about GPT-6 Astra being able to directly interpret model weights (prompts ...
精选理由
有人用 Qwen 3.8 27B 在笔记本上复现了 GPT-6 Astra 读权重的方法，跑了 45 分钟，过程和结论都写在博客里。
一篇新论文声称 GPT-6 Astra 能直接解读模型权重，即通过充满浮点数的提示词读取模型内部信息。Simon Willison 在自己的笔记本上用 Qwen 3.8 27B 复现了该实验。经过 45 分钟以上的运行，Qwen 3.8 27B 也实现了同样效果。
原文 · Simon Willison
There's a buzzy new paper about GPT-6 Astra being able to directly interpret model weights (prompts ...
There's a buzzy new paper about GPT-6 Astra being able to directly interpret model weights (prompts full of floating point numbers) - I tried it with Qwen 3.8 27B on my laptop and found that, after 45+ minutes, that model could achieve the same thing simonwillison.net/2026/Nov/3/qwe… 💬 55 🔄 13 ❤️ 141 👀 6202 📊 68 ⚡
