Zuchao Li

Protein as a Second Language for LLMs

TrigReason: Trigger-Based Collaboration between Small and Large Reasoning Models

RACER: Retrieval-Augmented Contextual Rapid Speculative Decoding

Turn-Level Trajectory Optimization for Robust Multi-Turn LLM Reasoning

SmallKV: Small Model Assisted Compensation of KV Cache Compression for Efficient LLM Inference

What Limits Bidirectional Model's Generative Capabilities? A Uni-Bi-Directional Mixture-of-Expert Method For Bidirectional Fine-tuning