Yilun Huang

New post-training method enhances LLM communication via convention formation

Researchers have developed a new post-training process to enhance the ability of large language models (LLMs) to form conventions and communicate more efficiently in multi-turn interactions. This fine-tuning method targ…