This paper explores the interaction modes between humans and Large Language Models (LLMs), particularly focusing on the interaction with ChatGPT. The authors identify four key phases in the human-LLM interaction flow: planning, facilitating, iterating, and testing. They also develop a taxonomy of four primary interaction modes: Standard Prompting, User Interface, Context-based, and Agent Facilitator. This taxonomy is refined using the "5W1H" guideline method, which examines definitions, participant roles, phases, objectives, and mechanics of each interaction mode. The paper aims to contribute to the design and evaluation of human-LLM interactions, particularly in complex tasks involving reasoning, creativity, and iteration. The authors conducted a systematic literature review of HCI papers published since 2021 to identify these interaction modes and phases. The taxonomy is expected to guide future research and applications in human-LLM interactions, including areas such as image and video generation. The paper discusses the potential applications of the taxonomy, such as enhancing team dynamics and performance through LLMs acting as facilitators, and highlights the need for further expansion and refinement of the taxonomy to include additional interaction modes and tasks.This paper explores the interaction modes between humans and Large Language Models (LLMs), particularly focusing on the interaction with ChatGPT. The authors identify four key phases in the human-LLM interaction flow: planning, facilitating, iterating, and testing. They also develop a taxonomy of four primary interaction modes: Standard Prompting, User Interface, Context-based, and Agent Facilitator. This taxonomy is refined using the "5W1H" guideline method, which examines definitions, participant roles, phases, objectives, and mechanics of each interaction mode. The paper aims to contribute to the design and evaluation of human-LLM interactions, particularly in complex tasks involving reasoning, creativity, and iteration. The authors conducted a systematic literature review of HCI papers published since 2021 to identify these interaction modes and phases. The taxonomy is expected to guide future research and applications in human-LLM interactions, including areas such as image and video generation. The paper discusses the potential applications of the taxonomy, such as enhancing team dynamics and performance through LLMs acting as facilitators, and highlights the need for further expansion and refinement of the taxonomy to include additional interaction modes and tasks.