Microsoft 研究:Claude Opus 5.5 直接手写 transformer 权重,无需梯度训练
这个思路很新颖:不训练模型,而是让 Claude Opus 5.5 直接把算法写成权重,166 个参数就解决了算法任务,人还能逐行读懂。相比传统梯度训练,可解释性是天然优势。做可解释性研究或科学建模的人会很有兴趣。
Microsoft 及合作团队发布了一项关于可解释权重空间编程(interpretable weight-space programming)的研究。该方法让 Claude Opus 5.5 在不做任何梯度训练的情况下,直接用代码写出 transformer 权重:智能体提出计算方案,将其写成显式的权重和特征,测试后在 30 到 50 轮迭代中反复修改。在算法类任务上,全部 26 次运行都在保留集上达到 100% 准确率,解决任务的最小模型参数中位数为 166,结果可被人类逐行阅读。
研究发现底层模型的选择影响显著:Claude Sonnet 5.5 完成算法任务所需的参数中位数为 41,600,而 Haiku 5.5 需要约 100,000。语言任务的难度明显更高,模型参数规模停留在约 190,000 附近,准确率低于通过梯度学习得到的基线。
团队还将同一套流程用于脑科学数据:在视觉皮层上,该方法的响应预测优于现有的可解释模型;在语言 fMRI 数据上,它发现了数据集中的一个混淆变量。这显示权重空间编程除了作为可解释性研究工具外,还能在科学建模场景中产生实际发现。
Super interesting work from Microsoft and colleagues.
They take a closer look at the exciting research area of weight-space programming.
They had Claude Opus 5.5 write transformer weights directly in code, with no gradient training.
The smallest models that solved the algorithmic tasks had a median of 166 parameters.
The method is called interpretable weight-space programming. An agent proposes a computation, writes it as explicit weights and features, tests it, and revises it for 30 to 50 rounds. A person can read the result line by line.
All 26 runs on algorithmic tasks reached 100% accuracy on held-out data, often by writing a known algorithm into the weights.
They find that the model matters. Sonnet 5.5 needed a median of 41,600 parameters, and Haiku 5.5 needed 100,000.
Language tasks were harder. Those models stayed near 190,000 parameters and lost accuracy against learned baselines.
They then used the same loop on brain data. In the visual cortex, it predicted responses better than existing interpretable models. In language fMRI, it found a confound in the dataset.
https://t.co/JKYdkdLqwV
Chat with Paper: https://t.co/N58d4lTGNC
本条情报归属 AITOP「论文」核心栏目,重点聚焦北美前沿实验室与硅谷科技巨头。在跨区域产业观察坐标中,该动态直接关联「AI 开发者工具链、Coding Agent 与工程化实践」研究主线,反映出当前技术演进与商业落地的最新风向。涉及核心实体与技术议题包括:Claude Opus 5.5、Microsoft、权重空间编程、可解释性,读者可通过下方标签进入对应专题页查阅历史脉络与关联对比。
信源核验说明:本文原始线索采编自「DAIR.AI」(@dair_ai),经 AITOP 多维评分管线结构化提炼。