Papers

• The Entropic Bound for Transformers: Why Static Rank Fails and Attention-Native Rank Recovers (2026)
  Byeong Hoon Yoon
  4.39
• A Coulomb Particle Model for Learning Kernel Attention in Transformers (2026)
  Masoud Badiei Khuzani et al.
  4.39
• Convergent Evolution: How Different Language Models Learn Similar Number Representations (2026)
  Deqing Fu et al.
  1.94
• SANA-WM: Efficient Minute-Scale World Modeling with Hybrid Linear Diffusion Transformer (2026)
  Haoyi Zhu et al.
  1.94
• Linear representations in language models can change dramatically over a conversation (2026)
  Andrew Kyle Lampinen et al.
  1.67
• Linear Correlation in LM's Compositional Generalization and Hallucination (2025)
  Letian Peng et al.
  1.28
• Quamba2: A Robust and Scalable Post-training Quantization Framework for Selective State Space Models (2025)
  Hung-Yueh Chiang et al.
  1.28
• It's All Connected: A Journey Through Test-Time Memorization, Attentional Bias, Retention, and Online Optimization (2025)
  Ali Behrouz et al.
  1.28
• Diagonal Batching Unlocks Parallelism in Recurrent Memory Transformers for Long Contexts (2025)
  Danil Sivtsov et al.
  1.28
• Discovering Hierarchical Latent Capabilities of Language Models via Causal Representation Learning (2025)
  Jikai Jin et al.
  1.28
• Ovis2.5 Technical Report (2025)
  Shiyin Lu et al.
  1.28
• SANA-Video: Efficient Video Generation with Block Linear Diffusion Transformer (2025)
  Junsong Chen et al.
  1.28
• Why Can't Transformers Learn Multiplication? Reverse-Engineering Reveals Long-Range Dependency Pitfalls (2025)
  Xiaoyan Bai et al.
  1.28
• Native Hybrid Attention for Efficient Sequence Modeling (2025)
  Jusen Du et al.
  1.28
• InfiniteVL: Synergizing Linear and Sparse Attention for Highly-Efficient, Unlimited-Input Vision-Language Models (2025)
  Hongyuan Tao et al.
  1.28
• Error-Free Linear Attention is a Free Lunch: Exact Solution from Continuous-Time Dynamics (2025)
  Jingdi Lei et al.
  1.28
• Mamba: Linear-Time Sequence Modeling with Selective State Spaces (2023)
  Albert Gu et al.
  —
• Knowledge Composition using Task Vectors with Learned Anisotropic Scaling (2024)
  Frederic Z. Zhang et al.
  —
• LinFusion: 1 GPU, 1 Minute, 16K Image (2024)
  Songhua Liu et al.
  —
• Cottention: Linear Transformers With Cosine Attention (2024)
  Gabriel Mongaras et al.
  —
• If You Can't Use Them, Recycle Them: Optimizing Merging at Scale Mitigates Performance Tradeoffs (2024)
  Muhammad Khalifa et al.
  —