0.6B追平32B，显存省50倍

重点关注

• Program-as-Weights: A Programming Paradigm for Fuzzy Functions score 10
  入选 HF Daily Papers；HF 热度: 81 upvotes (+4)；有代码实现；关键词(1): lightweight
• Multi-Resolution Flow Matching: Training-Free Diffusion Acceleration via Staged Sampling score 10
  入选 HF Daily Papers；HF 热度: 29 upvotes (+4)；有代码实现；关键词(3): lightweight, distillation, text-to-image
• AgenticSTS: A Bounded-Memory Testbed for Long-Horizon LLM Agents score 9
  入选 HF Daily Papers；HF 热度: 46 upvotes (+4)；有代码实现
• PACE: A Proxy for Agentic Capability Evaluation score 6
  入选 HF Daily Papers；HF 热度: 7 upvotes (+2)；关键词(3): agentic, code generation, reasoning
• EvoPolicyGym: Evaluating Autonomous Policy Evolution in Interactive Environments score 8
  入选 HF Daily Papers；HF 热度: 43 upvotes (+4)；关键词(1): leaderboard
• AGVBench: A Reliability-Oriented Benchmark of Data Augmentation for Vein Recognition score 8
  入选 HF Daily Papers；HF 热度: 13 upvotes (+3)；有代码实现
• AnyGroundBench: A Specialized-Domain Benchmark for Video Grounding in Vision-Language Models score 8
  入选 HF Daily Papers；HF 热度: 7 upvotes (+2)；有代码实现；关键词(3): pre-training, reasoning, vision-language
• SkillCoach: Self-Evolving Rubrics for Evaluating and Enhancing Agentic Skill-Use score 7
  入选 HF Daily Papers；HF 热度: 15 upvotes (+3)；关键词(1): agentic
• Denser $\neq$ Better: Limits of On-Policy Self-Distillation for Continual Post-Training score 8
  入选 HF Daily Papers；HF 热度: 5 upvotes (+2)；有代码实现；关键词(3): distillation, GRPO, post-training
• WorldDirector: Building Controllable World Simulators with Persistent Dynamic Memory score 7
  入选 HF Daily Papers；HF 热度: 21 upvotes (+4)

也值得关注

• Subliminal Clocks: Latent Time Modelling in Diffusion Language Models score 3
  机构: EPFL