VIABench: A Comprehensive Video Benchmark Collected from Blind Individuals for Visual Impairment Assistance Paper • 2607.14660 • Published 3 days ago • 7
The Mirage of Optimizing Training Policies: Monotonic Inference Policies as the Real Objective for LLM Reinforcement Learning Paper • 2606.29526 • Published 21 days ago • 167
StreamChar: Long-Horizon Streaming Character Audio-Video Generation with Decoupled Orchestration Paper • 2605.25659 • Published May 25 • 17
MASCing: Configurable Mixture-of-Experts Behavior via Activation Steering Masks Paper • 2604.27818 • Published Apr 30 • 5