A Taxonomy for Human-LLM Interaction Modes: An Initial Exploration

With ChatGPT's release, conversational prompting has become the most popular\nform of human-LLM interaction. However, its effectiveness is limited for more\ncomplex tasks involving reasoning, creativity, and iteration. Through a\nsystematic analysis of HCI papers published since 2021, we identified four key\nphases in the human-LLM interaction flow - planning, facilitating, iterating,\nand testing - to precisely understand the dynamics of this process.\nAdditionally, we have developed a taxonomy of four primary interaction modes:\nMode 1: Standard Prompting, Mode 2: User Interface, Mode 3: Context-based, and\nMode 4: Agent Facilitator. This taxonomy was further enriched using the "5W1H"\nguideline method, which involved a detailed examination of definitions,\nparticipant roles (Who), the phases that happened (When), human objectives and\nLLM abilities (What), and the mechanics of each interaction mode (How). We\nanticipate this taxonomy will contribute to the future design and evaluation of\nhuman-LLM interaction.\n