HF Daily Papers · 2026-11-02

HF Daily Papers 精选（15 篇，按社区票数排序）

  • [103▲] Look Only When Needed: Certificate-Guided Visual Routing for Efficient Diffusion VLMs — https://arxiv.org/abs/2610.36671
  • [72▲] A-RSI: Adaptive Task Refresh for Zero-Data Multimodal Self-Improvement — https://arxiv.org/abs/2610.38568
  • [66▲] Proving and Improving the Platonic Representation Hypothesis with JEPAs — https://arxiv.org/abs/2610.39302
  • [41▲] L3G-World: Real-Time Infinite World Generation with Latent 3D Gaussian Memory — https://arxiv.org/abs/2610.40614
  • [38▲] WorkFactory: Turning Professional Intents into Executable Work for Scalable Agent Training — https://arxiv.org/abs/2610.37161
  • [30▲] A Broader Look into MLLM's Retrieval Heads — https://arxiv.org/abs/2610.32411
  • [25▲] Read, Allocate, Route: Role-Coupled Tokenization for Pixel-Space Generative Transformers — https://arxiv.org/abs/2610.40647
  • [24▲] HairDeck: Animation-Ready Stylized Hair Cards from a Single Image — https://arxiv.org/abs/2610.40599
  • [24▲] VibeGame: Prompt-to-Game Development with AI-Native Engine and Self-Evolving Adversarial Agent Team — https://arxiv.org/abs/2610.33106
  • [22▲] Your Benchmark Is an Empirical Measure Over Difficulty — https://arxiv.org/abs/2610.29927
  • [21▲] Co-Design Bench: Multimodal Collaborative Coding Evaluation for Agentic Models — https://arxiv.org/abs/2610.35138
  • [20▲] Intrinsic Distribution Matching — https://arxiv.org/abs/2610.39633
  • [19▲] WeVisCut: The Whole Editing Room in One Agent From Narration to Long-Form Video — https://arxiv.org/abs/2610.38352
  • [18▲] Multi-Token Prediction Done Right — https://arxiv.org/abs/2610.38331
  • [18▲] RealCLI: Toward a Unified Evaluation of Human-Agent Collaboration for CLI Agents — https://arxiv.org/abs/2610.22507