Hongyin Luo

Beyond Context Limits: Subconscious Threads for Long-Horizon Reasoning

ROVER: Recursive Reasoning Over Videos with Vision-Language Models for Embodied Tasks

Quantifying Generalization Complexity for Large Language Models

Self-MoE: Towards Compositional Large Language Models with Self-Specialized Experts

Addition is All You Need for Energy-efficient Language Models

Training Task Experts through Retrieval Based Distillation