创建

登录 ReadmeX

或

了解我们如何处理你的信息: 隐私政策

ReadmeX

陪你读懂 AI 动态

今天,想先了解什么?

选一份动态总结,或直接聊聊你的问题。

社区和关注动态总结只属于你自己。

资讯

Claude Fable 5.2实测MNIST直接权重编写

AI 总结

在一项由单一作者开展的实验中,Claude Fable 5.2生成代码,从零设置了一个固定MNIST卷积神经网络的全部5,994个权重。规则禁止拟合、回归和搜索,但允许手工绘制滤波器与模板;实验也不得使用PyTorch或进行练习对局,每次尝试后只向模型提供训练损失、准确率和混淆矩阵。Claude最多可执行50步,在此过程中逐步完善手工设计的特征、测量特征在训练集中的出现频率,并将最终层重建为朴素贝叶斯式分类器。网络准确率达到约90%。作者认为,这一结果是对历史性手工建模技术的组合运用,而非一种全新的训练方法,并将其与GPT-6 Astra的围棋网络进行比较。

深度见解

该实验检验了LLM直接编写并迭代修改神经网络权重的可能性,同时表明其报告结果依赖手工设计的特征和对训练数据的测量,而非传统的自动优化流程。

The writing aspect of WSP turns out to be quite a history lesson.

Image 1

Intro

Weight-Space Programming still has me pretty intrigued. The basic idea is that frontier LLMs can read, write and edit the weights of neural networks. There's nuance to what's going on, but it feels like there's a lot of new ground to explore.

In the original blog post, beyond the reading and editing experiments, the authors had GPT-6 Astra build a network to play Go from scratch, then improve it by playing practice games, and editing the weights based on analysis of its losses. That loop of writing, testing and editing feels like a new way to train a network, something like LLM Descent. But is it?

In this article, I run my own weight-writing experiment with Claude Fable 5.2 on a familiar image classification problem. Look closely at what the model builds, and it turns out to be an interesting lesson in historical model-building techniques. With that new lens on weight-space programming, I also study the Go network that Astra built in the original article, and find something very similar.

#Claude Fable 5.2#GPT-6 Astra#Weight-Space Programming#MNIST

原文约 16 分钟读完阅读原文
群聊

还没有评论,来说说你的看法。