Wulong Liu

Principled Fast and Meta Knowledge Learners for Continual Reinforcement Learning

TrimR: Verifier-based Training-Free Thinking Trimming for Efficient Test-Time Scaling

Analytical Restructuring of Feed-Forward Networks for Accelerated LLM Inference

FlatQuant: Flatness Matters for LLM Quantization

Principled Fast and Meta Knowledge Learners for Continual Reinforcement Learning

AttentionPredictor: Temporal Patterns Matter for KV Cache Compression

FlatQuant: Flatness Matters for LLM Quantization

KVTuner: Sensitivity-Aware Layer-Wise Mixed-Precision KV Cache Quantization for Efficient and Nearly Lossless LLM Inference