GRAPHMOE: Amplifying Cognitive Depth of Mixture-of-Experts Network via Introducing Self-Rethinking Mechanism Paper • 2501.07890 • Published Jan 14, 2025 • 1
Innovator: Scientific Continued Pretraining with Fine-grained MoE Upcycling Paper • 2507.18671 • Published Jul 24, 2025 • 1
ssToken: Self-modulated and Semantic-aware Token Selection for LLM Fine-tuning Paper • 2510.18250 • Published Oct 21, 2025 • 14
Repulsor: Accelerating Generative Modeling with a Contrastive Memory Bank Paper • 2512.08648 • Published Dec 9, 2025
HELP: HyperNode Expansion and Logical Path-Guided Evidence Localization for Accurate and Efficient GraphRAG Paper • 2602.20926 • Published 17 days ago • 3
DreamWorld: Unified World Modeling in Video Generation Paper • 2603.00466 • Published 14 days ago • 16
GRADE: Benchmarking Discipline-Informed Reasoning in Image Editing Paper • 2603.12264 • Published 1 day ago • 13
EvoTok: A Unified Image Tokenizer via Residual Latent Evolution for Visual Understanding and Generation Paper • 2603.12108 • Published 1 day ago • 8