叶子的技术碎碎念

Lumine: An Open Recipe for Building Generalist Agents in 3D Open Worlds Lumine: 构建3D开放世界中通用AI智能体的开放方

Thinking with Video: Video Generation as a Promising Multimodal Reasoning Paradigm 视频思维：视频生成作为一种前景广阔

Scaling Latent Reasoning via Looped Language Models 通过循环语言模型扩展潜在推理 现代大语言模型主要通过显式文本生成（如思维链 (CoT)）来训练其

A Theoretical Study on Bridging Internal Probability and Self-Consistency for LLM Reasoning 关于桥接大语言模

QeRL：超越效率——面向大语言模型的量化增强强化学习 QeRL：超越效率——面向大语言模型的量化增强强化学习 我们提出 QeRL，一种面向大语言模型 (LLMs) 的量化增强强化学习框架。尽管强化学

Less is More: Recursive Reasoning with Tiny Networks 少即是多：微型网络的递归推理 分层推理模型 (HRM) 是一种创新方法，它采用两个小型神经网络

The Dragon Hatchling: The Missing Link between the Transformer and Models of the Brain 龙之雏形：Transfor

OmniWorld: A Multi-Domain and Multi-Modal Dataset for 4D World Modeling OmniWorld: A Multi-Domain an

Sharing is Caring: Efficient LM Post-Training with Collective RL Experience Sharing 分享即关怀：通过集体强化学习经验