Lattice AI Research

Research focus

Architecture Design (Transformers, SSMs, MoE) (1)Training Efficiency & Optimization (1)Interpretability & Mechanistic Interp (1)Natural Language Processing (1)

Frequent co-authors

Teng Pang (1)Zhiqiang Dong (1)Yan Zhang (1)Rongjian Xu (1)

Papers (2)

Apr 9, 2026

1w ago

Value-Guidance MeanFlow for Offline Multi-Agent Reinforcement Learning

Ditch the multi-step sampling and regularization coefficient tuning: VGM$^2$P achieves SOTA offline MARL performance with a simple, efficient flow-based policy guided by global advantage values.

Teng Pang, Zhiqiang Dong, Yan Zhang +4

Architecture Design (Transformers, SSMs, MoE)Training Efficiency & Optimization

1w ago·also Salesforce AI

Distributed Multi-Layer Editing for Rule-Level Knowledge in Large Language Models

Editing a rule in an LLM is not like editing a fact; you can't just tweak one layer – formulas live in early layers, instances in the middle.

Yating Wang, Wenting Zhao, Yaqi Zhao +4

Interpretability & Mechanistic Interp Natural Language Processing Reasoning & Chain-of-Thought

Yilong Yin

Publication activitypapers/week, last 8 weeks

Research focus

Frequent co-authors

Papers (2)

Search

Yilong Yin

Publication activitypapers/week, last 8 weeks

Research focus

Frequent co-authors

Papers (2)