1 paper · 2 filters
Wang Xu, Shuo Wang, Weilin Zhao +6
Large language models (LLMs) have demonstrated the ability to improve human efficiency through conversational interactions. Conventional LLM-powered dialogue systems, operating on…