AI前沿：从图像生成到数学推理

[LG] Guidance in the Frequency Domain Enables High-Fidelity Sampling at Low CFG Scales

[ETH Zürich & DisneyResearch|Studios]

arxiv.org

---

[LG] Thought Anchors: Which LLM Reasoning Steps Matter?

[Duke University & Aiphabet]

arxiv.org

---

[CL] LongWriter-Zero: Mastering Ultra-Long Text Generation via Reinforcement Learning

[Singapore University of Technology and Design & Tsinghua University]

arxiv.org

---

[CL] OMEGA: Can LLMs Reason Outside the Box in Math? Evaluating Exploratory, Compositional, and Transformative Generalization

[dmodel.ai & UC Berkeley]

arxiv.org