[LG] Is your batch size the problem? Revisiting the Adam-SGD gap in language modeling  

[Max Planck Institute for Intelligent Systems]  

https://arxiv.org/abs/2506.12543

你会发现，决定你五年、十年后成为一个什么样的人，根本不是你那些写在纸上、存在手机里的宏伟计划，而是你今天，此时此刻，选择拿起一本书，还是刷一个短视频；是选择穿上跑鞋，还是继续窝在沙发里……

[LG] Because we have LLMs, we Can and Should Pursue Agentic Interpretability

[Google DeepMind]

https://arxiv.org/abs/2506.12152

[LG] Text-to-LoRA: Instant Transformer Adaption

[Sakana AI]

https://arxiv.org/abs/2506.06105

[LG] On the Closed-Form of Flow Matching: Generalization Does Not Arise from Target Stochasticity

[CNRS]

https://arxiv.org/abs/2506.03719

[CL] Don't Pay Attention

[Avey AI]

https://arxiv.org/abs/2506.11305

真正的知识管理，不是收藏，而是创造。

[LG] AutoRule: Reasoning Chain-of-thought Extracted Rule-based Rewards Improve Preference Learning

[CMU]

https://arxiv.org/abs/2506.15651

[LG] HeurAgenix: Leveraging LLMs for Solving Complex Combinatorial Optimization Challenges

[Microsoft Research Asia]

https://arxiv.org/abs/2506.15196

[LG] Dense SAE Latents Are Features, Not Bugs

[MIT & ETH Zürich]

https://arxiv.org/abs/2506.156