AI可可AI生活

[CL] Bridging Offline and Online Reinforcement Learning for LLMs [FAIR at Meta] arxiv.org 在小宇宙查看该单集文稿

[LG] The Ideation-Execution Gap: Execution Outcomes of LLM-Generated versus Human Research Ideas [Stanford University] arxiv.org 在小宇宙查看该单集文稿

[CL] Potemkin Understanding in Large Language Models [MIT & University of Chicago & Harvard University] arxiv.org 在小宇宙查看该单集文稿

[CL] Can Gradient Descent Simulate Prompting? [MIT CSAIL] arxiv.org 在小宇宙查看该单集文稿

[CL] OctoThinker: Mid-training Incentivizes Reinforcement Learning Scaling [Shanghai Jiao Tong University] https://arxiv.org/abs/2506.20512 --- [LG] Overtuning in Hyperparameter Optimization [LMU Munich] https://arxiv.org/abs/2506.19540 --- [LG] Distilling Normalizing Flows [University of Oregon & HSE University & Picsart AI Research] https://arxiv.org/abs/2506.21003 --- [LG] Gaussian Invariant Ma...

这个时代对普通人最底层的要求，其实就是做一个“正常人”。 在小宇宙查看该单集文稿

[LG] Asymmetric REINFORCE for off-Policy Reinforcement Learning: Balancing positive and negative rewards   [FAIR at Meta]   arxiv.org 在小宇宙查看该单集文稿

[LG] Mastering Multiple-Expert Routing: Realizable H-Consistency and Strong Guarantees for Learning to Defer   [Courant Institute of Mathematical Sciences & Google Research]   arxiv.org 在小宇宙查看该单集文稿

[CL] Inside you are many wolves: Using cognitive models to interpret value trade-offs in LLMs   [Harvard University]   arxiv.org 在小宇宙查看该单集文稿

[LG] Language Modeling by Language Models   [Allen Institute for AI]   arxiv.org 在小宇宙查看该单集文稿

[CL] DiffuCoder：Understanding and Improving Masked Diffusion Models for Code Generation   [Apple]   arxiv.org 在小宇宙查看该单集文稿

[CL] Can Gradient Descent Simulate Prompting? [MIT CSAIL] https://arxiv.org/abs/2506.20989 --- [CL] Potemkin Understanding in Large Language Models [MIT & University of Chicago & Harvard University] https://arxiv.org/abs/2506.21521 --- [LG] The Ideation-Execution Gap: Execution Outcomes of LLM-Generated versus Human Research Ideas [Stanford University] https://arxiv.org/abs/2506.20803 --- [CL] Bri...

决定你人生品质的，不是世界怎么看你，而是你如何看待你自己的世界。 在小宇宙查看该单集文稿

[LG] Orthogonal Finetuning Made Scalable [Max Planck Institute for Intelligent Systems & University of Cambridge] arxiv.org 在小宇宙查看该单集文稿

[LG] Who Does What in Deep Learning? Multidimensional Game-Theoretic Attribution of Function of Neural Units [University Medical Center Eppendorf & University of Cambridge] https://arxiv.org/abs/2506.19732 在小宇宙查看该单集文稿

[LG] Command-V: Pasting LLM Behaviors via Activation Profiles [CMU] https://arxiv.org/abs/2506.19140 在小宇宙查看该单集文稿

[CL] Broken Tokens? Your Language Model can Secretly Handle Non-Canonical Tokenizations [University of Washington] https://arxiv.org/abs/2506.19004 在小宇宙查看该单集文稿

[LG] Chain-of-Experts: Unlocking the Communication Power of Mixture-of-Experts Models [Northwestern University] https://arxiv.org/abs/2506.18945 在小宇宙查看该单集文稿

[CL] DiffuCoder：Understanding and Improving Masked Diffusion Models for Code Generation [Apple] https://arxiv.org/abs/2506.20639 --- [LG] Language Modeling by Language Models [Allen Institute for AI] https://arxiv.org/abs/2506.20249 --- [CL] Inside you are many wolves: Using cognitive models to interpret value trade-offs in LLMs [Harvard University] https://arxiv.org/abs/2506.20666 --- [LG] Maste...

这个世界上最大的“监狱”，不是用钢筋水泥建造的，而是我们自己思维的产物。 在小宇宙查看该单集文稿

[CL] Cite Pretrain: Retrieval-Free Knowledge Attribution for Large Language Models [Duke University & Meta] https://arxiv.org/abs/2506.17585 在小宇宙查看该单集文稿

[LG] In-Context Learning Strategies Emerge Rationally [Stanford University & Harvard University] https://arxiv.org/abs/2506.17859 在小宇宙查看该单集文稿

[LG] Routing Mamba: Scaling State Space Models with Mixture-of-Experts Projection [Microsoft] https://arxiv.org/abs/2506.18145 在小宇宙查看该单集文稿

[LG] The 4th Dimension for Scaling Model Size [University of Illinois at Urbana-Champaign & University of Toronto] https://arxiv.org/abs/2506.18233 在小宇宙查看该单集文稿

[LG] These are Not All the Features You are Looking For: A Fundamental Bottleneck In Supervised Pretraining [Facebook AI Research (FAIR) at Meta & New York University] https://arxiv.org/abs/2506.18221 在小宇宙查看该单集文稿