Hierarchical Sparse Attention Done Right: Toward Infinite Context Modeling Paper • 2607.02980 • Published 30 days ago • 83
Locas: Your Models are Principled Initializers of Locally-Supported Parametric Memories Paper • 2602.05085 • Published Feb 4 • 5
Can LLMs Guide Their Own Exploration? Gradient-Guided Reinforcement Learning for LLM Reasoning Paper • 2512.15687 • Published Dec 17, 2025 • 22