Thinking Inertia: LLMs Keep Thinking When Told Not To Paper • 2610.11765 • Published 6 days ago • 215
SepGen: Multi-Stem Audio-Video Separation and Generation in a Single Model Paper • 2610.11361 • Published 6 days ago • 39
Memento 3: Model-Based Recursive Self-Improvement through Reflective Rulebooks Paper • 2610.11794 • Published 6 days ago • 42
DeepSeek-V4.1-Flash: Pushing the Limits of KV Cache Compression Paper • 2609.19969 • Published 27 days ago • 229
An Open Recipe for IMO Gold: Training Nemotron for Olympiad Mathematics Paper • 2609.10712 • Published Sep 9 • 45
Recognition-Refusal Misalignment in LLMs: Why Models Answer Structurally Unanswerable Questions Paper • 2608.29109 • Published Aug 29 • 17
Knowing When Not to Reuse: Conditional Experience Transfer in Autonomous LLM Post-Training Paper • 2608.26730 • Published Aug 27 • 154