Daily Papers of 2025-07-11

  1. Scaling RL to Long Videos 132 upvotes, #1 of 2025-07-11
  2. T-LoRA: Single Image Diffusion Model Customization Without Overfitting 105 upvotes, #2 of 2025-07-11
  3. Multi-Granular Spatio-Temporal Token Merging for Training-Free Acceleration of Video LLMs 44 upvotes, #3 of 2025-07-11
  4. Traceable Evidence Enhanced Visual Grounded Reasoning: Evaluation and Methodology 44 upvotes, #3 of 2025-07-11
  5. OST-Bench: Evaluating the Capabilities of MLLMs in Online Spatio-temporal Scene Understanding 37 upvotes, #5 of 2025-07-11
  6. LangSplatV2: High-dimensional 3D Language Gaussian Splatting with 450+ FPS 32 upvotes, #6 of 2025-07-11
  7. Geometry Forcing: Marrying Video Diffusion and 3D Representation for Consistent World Modeling 30 upvotes, #7 of 2025-07-11
  8. Skip a Layer or Loop it? Test-Time Depth Adaptation of Pretrained LLMs 28 upvotes, #8 of 2025-07-11
  9. PyVision: Agentic Vision with Dynamic Tooling 28 upvotes, #8 of 2025-07-11
  10. A Survey on Long-Video Storytelling Generation: Architectures, Consistency, and Cinematic Quality 22 upvotes, #10 of 2025-07-11
  11. Token Bottleneck: One Token to Remember Dynamics 17 upvotes, #11 of 2025-07-11
  12. Dynamic Chunking for End-to-End Hierarchical Sequence Modeling 17 upvotes, #11 of 2025-07-11
  13. Machine Bullshit: Characterizing the Emergent Disregard for Truth in Large Language Models 16 upvotes, #13 of 2025-07-11
  14. SciMaster: Towards General-Purpose Scientific AI Agents, Part I. X-Master as Foundation: Can We Lead on Humanity's Last Exam? 4 upvotes, #14 of 2025-07-11
  15. Beyond the Linear Separability Ceiling 4 upvotes, #14 of 2025-07-11
  16. Emergent Semantics Beyond Token Embeddings: Transformer LMs with Frozen Visual Unicode Representations 2 upvotes, #16 of 2025-07-11
  17. Growing Transformers: Modular Composition and Layer-wise Expansion on a Frozen Substrate 2 upvotes, #16 of 2025-07-11
  18. Re-Bottleneck: Latent Re-Structuring for Neural Audio Autoencoders 2 upvotes, #16 of 2025-07-11

Data: hysts-bot-data/daily-papers-stats and the Daily Papers API. Open data: tardellirs/paper-pulse-data. Sister project: Model Pulse, the download history of every model on the Hub.