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。

Task Vectors [in Vision and Language Models (VLM's)] Are Cross-Modal

1 点作者 peter_d_sherman6 个月前

1 comment

peter_d_sherman6 个月前
&gt;&quot;In the in-context learning (ICL) paradigm, given a set of examples, the model has to learn the mapping from inputs to outputs. Prior research has demonstrated that LLMs implicitly compress this mapping into a latent activation, called the <i>task vector</i>&quot;<p>Related:<p>In-Context Learning Creates Task Vectors:<p><a href="https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;2310.15916" rel="nofollow">https:&#x2F;&#x2F;arxiv.org&#x2F;abs&#x2F;2310.15916</a><p>Function Vectors in Large Language Models:<p><a href="https:&#x2F;&#x2F;functions.baulab.info&#x2F;" rel="nofollow">https:&#x2F;&#x2F;functions.baulab.info&#x2F;</a><p><a href="https:&#x2F;&#x2F;x.com&#x2F;graceluo_&#x2F;status&#x2F;1852043048043319360" rel="nofollow">https:&#x2F;&#x2F;x.com&#x2F;graceluo_&#x2F;status&#x2F;1852043048043319360</a>