Yu, Le
10 publications
NeurIPS
2025
Beyond the 80/20 Rule: High-Entropy Minority Tokens Drive Effective Reinforcement Learning for LLM Reasoning
NeurIPS
2025
Gated Attention for Large Language Models: Non-Linearity, Sparsity, and Attention-Sink-Free
AAAI
2025
Revolutionizing Encrypted Traffic Classification with MH-Net: A Multi-View Heterogeneous Graph Model