21IC电子网

DeepSeek发布梁文锋署名新论文 V4有望支持全新记忆架构

1月13日消息，今日，DeepSeek发布新论文《Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models》 (基于可扩展查找的条件记忆：大型语言模型稀疏性的新维度)。

DeepSeek发布梁文锋署名新论文，开源相关记忆模块Engram

DeepSeek于12日晚发布新论文《Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language ...

刚刚，梁文锋署名开源「记忆」模块，DeepSeek V4更细节了

就在十几个小时前，DeepSeek 发布了一篇新论文，主题为《Conditional Memory via Scalable Lookup:A New Axis of Sparsity for Large Language ...

Unlocking Business Value With Open-Weight Large Language Models

Open-weight LLMs can unlock significant strategic advantages, delivering customization and independence in an increasingly AI ...

AI researchers are now studying LLMs as if they were living organisms

Large language models have grown so vast and complex that even the people who build them no longer fully understand how they work. A single modern ...

DeepSeek发布梁文锋署名新论文 开源记忆模块

2026年1月13日，DeepSeek与北京大学合作发布新论文《Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models》，创始人梁文锋为合著作者之一。论文提出条件记忆（conditional memory）概念，通过可扩展查找结构解决大语言模型知识检索效率低下的问题。同日，团队 ...
Forbes

Large Behavior Models Surpass Large Language Models To Create AI That Walks And Talks

Forbes contributors publish independent expert analyses and insights. Dr. Lance B. Eliot is a world-renowned AI scientist and consultant. In today’s column, I closely explore the rapidly emerging ...
Semiconductor Engineering

Small Vs. Large Language Models

The proliferation of edge AI will require fundamental changes in language models and chip architectures to make inferencing and learning outside of AI data centers a viable option. The initial goal ...

Autocomplete: Large language models can repeat training data verbatim

Researchers show that LLMs can reproduce copyrighted training data almost verbatim. This means headaches for model providers.