A curated digest of arXiv and Hugging Face papers that caught my eye — mostly around LLM agents, reasoning, and reinforcement learning for LLMs. Collected automatically and updated regularly.