MOSS-Audio Collection An open-source audio understanding model supporting speech recognition, environmental sound analysis, music understanding, time-aware QA, and complex • 5 items • Updated about 16 hours ago • 35
The Past Is Not Past: Memory-Enhanced Dynamic Reward Shaping Paper • 2604.11297 • Published 4 days ago • 132
BandPO: Bridging Trust Regions and Ratio Clipping via Probability-Aware Bounds for LLM Reinforcement Learning Paper • 2603.04918 • Published Mar 5 • 56
EvoCUA: Evolving Computer Use Agents via Learning from Scalable Synthetic Experience Paper • 2601.15876 • Published Jan 22 • 92
HERMES: KV Cache as Hierarchical Memory for Efficient Streaming Video Understanding Paper • 2601.14724 • Published Jan 21 • 75
FutureOmni: Evaluating Future Forecasting from Omni-Modal Context for Multimodal LLMs Paper • 2601.13836 • Published Jan 20 • 37