arXiv cs.AI· Yash Mehta, Chandan Singh, Weijia Xu, Richard J. Antonello, Adeen Flinker, Nima Mesgarani, Michael F. Bonner, Nebojsa Jojic, Jianfeng Gao·· 4 小时前AI 评分54
iWSP 方法用 LLM 智能体直接编写权重构建可解释 Transformer 与脑编码模型
Interpretable weight-space programming builds transparent transformers and brain encoding models
AI 导读
论文提出可解释权重空间编程(iWSP),让 LLM 智能体不依赖反向传播,以代码显式编写网络权重与特征,迭代构建可解释模型。作者先用算术与情感分类等基准验证方法,再构建覆盖视觉、语言、听觉的神经编码模型,在视觉皮层预测优于现有可解释基线,并在听觉 ECoG 中系统性发现可解释声学特征。
正文 · 原文
Abstract:Neural networks are powerful predictors, yet their opacity limits their usefulness in science, where the goal is explanation rather than mere prediction. In neuroscience, for example, deep encoding models accurately predict brain activity in response to natural stimuli, but struggle to reveal which stimulus features drive those predictions. Here we introduce interpretable weight-space programming (iWSP), a method in which an LLM agent iteratively builds interpretable models by explicitly programming their weights and features in code, without gradient-based training (backpropagation). Without backpropagation to tune parameters, the agent must draw on its prior knowledge and reason over previous trials to hypothesize, hand-craft, and refine human-interpretable computations mirroring the workflow of a human scientist. LLM audits confirm the interpretability of the hand-crafted models. We first validate iWSP by tasking it with constructing transformers for canonical benchmarks such as arithmetic and sentiment classification, finding that LLM agents can successfully craft compact, accurate networks directly. Next, we apply iWSP to build neural encoding models across vision, language, and audition. In the visual cortex, iWSP outpredicts existing interpretable baselines, uncovers distinct computations across early/higher visual areas, and generalizes to unseen data. In language fMRI, iWSP surfaces a confound in the data and in auditory ECoG, it systematically discovers interpretable acoustic features that predict neural activity. Together, these results establish iWSP as a general paradigm for aligning high predictive performance with human-interpretable scientific discovery.
| Subjects: | Computational Engineering, Finance, and Science (cs.CE); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.18114 [cs.CE] |
| (or arXiv:2611.18114v1 [cs.CE] for this version) | |
| https://doi.org/10.48550/arXiv.2611.18114 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Yash Mehta [view email]
[v1]
Fri, 13 Nov 2026 12:36:05 UTC (3,177 KB)
来源:arXiv cs.AI · arxiv.org