The Anh Han

LLMs may exhibit strong rule recall but often fail to track game states effectively, revealing critical gaps in their strategic reasoning.

A reputation-based mechanism can elevate cooperation rates in decentralized federated learning from below 5% to nearly 100%, drastically improving accuracy and stability.

Competitive pressure can drive developers to make unsafe choices, revealing that fear of falling behind is a stronger motivator than individual risk preferences.

Medical QA accuracy jumps to 87% when combining specialized LLM agents for reasoning, evidence retrieval, and refinement, demonstrating a practical path to more reliable clinical AI.