This paper presents a taxonomy for human-LLM interaction modes, identifying four key phases in the interaction flow: planning, facilitating, iterating, and testing. It also introduces four primary interaction modes: Mode 1: Standard Prompting, Mode 2: User Interface, Mode 3: Context-based, and Mode 4: Agent Facilitator. The taxonomy was developed through a systematic review of HCI literature published since 2021, using the "5W1H" guideline method to define the phases, roles, objectives, and mechanics of each interaction mode. The taxonomy aims to provide a structured framework for understanding and evaluating human-LLM interactions, enabling more effective design and implementation of interaction modes. The research highlights the importance of considering various interaction perspectives and the potential for future work in expanding the taxonomy to include different types of tasks and design spaces. The study also discusses the limitations of the current taxonomy and suggests future research directions. The findings contribute to the broader understanding of human-LLM interaction and offer a foundation for further exploration in this area.This paper presents a taxonomy for human-LLM interaction modes, identifying four key phases in the interaction flow: planning, facilitating, iterating, and testing. It also introduces four primary interaction modes: Mode 1: Standard Prompting, Mode 2: User Interface, Mode 3: Context-based, and Mode 4: Agent Facilitator. The taxonomy was developed through a systematic review of HCI literature published since 2021, using the "5W1H" guideline method to define the phases, roles, objectives, and mechanics of each interaction mode. The taxonomy aims to provide a structured framework for understanding and evaluating human-LLM interactions, enabling more effective design and implementation of interaction modes. The research highlights the importance of considering various interaction perspectives and the potential for future work in expanding the taxonomy to include different types of tasks and design spaces. The study also discusses the limitations of the current taxonomy and suggests future research directions. The findings contribute to the broader understanding of human-LLM interaction and offer a foundation for further exploration in this area.