Pingzhi Li

Adaptive Thinking: Large Language Models Know When to Think in Latent Space

Graph-of-Agents: A Graph-based Framework for Multi-Agent LLM Collaboration

TRUST: A Decentralized Framework for Auditing Large Language Model Reasoning

AsyncSpade: Efficient Test-Time Scaling with Asynchronous Sparse Decoding

DOGe: Defensive Output Generation for LLM Protection Against Knowledge Distillation

AlphaFormer: End-to-End Symbolic Regression of Alpha Factors with Transformers

Make Optimization Once and for All with Fine-grained Guidance

Can GRPO Help LLMs Transcend Their Pretraining Origin?

Mozart: Modularized and Efficient MoE Training on 3.5D Wafer-Scale Chiplet Architectures

Occult: Optimizing Collaborative Communications across Experts for Accelerated Parallel MoE Training and Inference

PortLLM: Personalizing Evolving Large Language Models with Training-Free and Portable Model Patches

Understanding Prejudice and Fidelity of Diverge-to-Converge Multi-Agent Systems

Examining Post-Training Quantization for Mixture-of-Experts: A Benchmark

Fantastic Experts and How to Find Them: A Multi-Dimensional Study for Experts-Level Sparsification in Mixture-of-Experts

Glider: Global and Local Instruction-Driven Expert Router