TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

New technique probes 'functional networks' in LLM with fMRI statistical analysis

2 点作者 cavilling_elite超过 1 年前

1 comment

cavilling_elite超过 1 年前
Researchers have adapted techniques from functional neuroimaging to map internal &quot;functional networks&quot; in large language models. By presenting the model with task-based prompts in an on&#x2F;off pattern, they identified distinct networks preferentially activated for domains like political science, radiology, and paleontology.<p>They found the greatest overlap between networks for radiology and pathology, suggesting shared semantic representations. The activated connections were fairly consistent across runs with similar tasks, and strongly predicted the prompted task in held-out experiments.<p>This offers a new way to peel back the black box of large models and understand their hidden organizational logic. It could help explain failures, selectively target parts of a model for fine-tuning, or monitor &quot;alignment&quot; during inference. The techniques parallel those used for decades to map functional networks in the brain.