记忆让agent谄媚，视觉推理93.2%

重点关注

• Seed2.0 Model Card: Towards Intelligence Frontier for Real-World Complexity score 11
  机构: ByteDance；入选 HF Daily Papers；HF 热度: 23 upvotes (+4)；关键词(1): reasoning
• MemSyco-Bench: Benchmarking Sycophancy in Agent Memory score 10
  入选 HF Daily Papers；HF 热度: 22 upvotes (+4)；有代码实现；关键词(1): reasoning
• VideoSearch-R1: Iterative Video Retrieval and Reasoning via Soft Query Refinement score 9
  入选 HF Daily Papers；HF 热度: 16 upvotes (+3)；有代码实现；关键词(3): GRPO, agentic, reasoning
• Perceive-to-Reason: Decoupling Perception and Reasoning for Fine-Grained Visual Reasoning score 9
  入选 HF Daily Papers；HF 热度: 13 upvotes (+3)；有代码实现；关键词(3): GRPO, reasoning, vision-language
• Valdi: Value Diffusion World Models score 9
  入选 HF Daily Papers；HF 热度: 10 upvotes (+3)；有代码实现；关键词(1): latency
• Multimodal Continuous Reasoning via Asymmetric Mutual Variational Learning score 8
  入选 HF Daily Papers；HF 热度: 22 upvotes (+4)；关键词(1): reasoning
• ABot-M0.5: Unified Mobility-and-Manipulation World Action Model score 7
  入选 HF Daily Papers；HF 热度: 13 upvotes (+3)；关键词(1): embodied
• Graph-Native Reinforcement Learning Enables Traceable Scientific Hypothesis Generation through Conceptual Recombination score 8
  入选 HF Daily Papers；HF 热度: 9 upvotes (+2)；有代码实现；关键词(2): GRPO, reasoning
• PixelEyes: Decoupling Perception and Reasoning for Pinpoint Visual Evidence Seeking score 8
  入选 HF Daily Papers；HF 热度: 7 upvotes (+2)；有代码实现；关键词(1): reasoning
• CausalMix: Data Mixture as Causal Inference for Language Model Training score 6
  入选 HF Daily Papers；HF 热度: 17 upvotes (+3)

也值得关注

• Multi-Turn Agentic Scientific Literature Search via Workflow Induction score 4
  机构: Stanford；关键词(2): agentic, reasoning