Simon Willison:本地 Qwen 3.8 27B 经过 45 分钟也能解读模型权重
中文全文 · AI 翻译
最近有篇新论文引起热议,称 GPT-6 Astra 能直接解读模型权重(也就是由大量浮点数组成的提示词)。我在笔记本电脑上用 Qwen 3.8 27B 试了试,发现运行 45 分钟以上后,这个模型也能做到同样的事:https://t.co/6O7WltGP7K
对照原文
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 https://t.co/6O7WltGP7K