Prompt
Let's come up with a 20 part taxonomy for classifying typical LLM interactions. Let's divide them according to whether they come from the user or the LLM. Create 10 for each.
Output
Here's a 20-part taxonomy for classifying typical LLM (Large Language Model) interactions, divided into user-driven and LLM-driven components:
User-Driven Interactions
- Prompt:
- The initial query or instruction given by the user to the LLM.
- Clarification Request:
- When the user asks for more detailed information or clarification on a previous response.
- Follow-up Question:
- A related question that builds on the LLM’s previous response.
- Feedback:
- Explicit feedback from the user, such as agreeing, disagreeing, or providing corrections.
- Contextual Information:
- Additional background or context provided by the user to guide the LLM’s response.
- Task Instruction:
- A command or directive given by the user, such as "Translate this" or "Summarize the text."
- Preference Indication:
- The user specifies a preference, such as tone, format, or style for the response.
- Error Report:
- The user indicates that the LLM made an error, prompting correction or revision.
- Example Provision:
- The user provides an example to help the LLM understand the task or question better.
- Session Closure:
- The user indicates that the interaction is complete, signaling the end of the session.
LLM-Driven Interactions
- Initial Response:
- The LLM’s first answer or output generated in response to the user’s prompt.
- Clarification Query:
- The LLM asks the user for more information to better understand the prompt.
- Contextual Continuation:
- The LLM maintains context across multiple turns in a conversation, continuing from previous responses.
- Suggestion:
- The LLM proposes ideas, alternative approaches, or additional information that might be relevant.
- Error Acknowledgment:
- The LLM recognizes a mistake in its response and offers a correction.
- Response Expansion:
- The LLM provides an elaborated or extended response, going beyond what was asked.
- Hypothetical Scenarios:
- The LLM generates hypothetical examples or scenarios to illustrate a point or answer a question.
- Assumption Making:
- The LLM makes assumptions based on incomplete information and provides a response accordingly.
- Tone Adaptation:
- The LLM adjusts the tone, style, or formality of its response based on the user’s input or preferences.
- Session Continuation:
- The LLM prompts the user to continue the conversation or explore related topics.
This taxonomy helps classify and analyze interactions, providing a structured way to understand the dynamics of communication between users and LLMs.## Extracted Prompts Prompts/Autocapture/LLM Interaction Classification.md#Prompt 1
Extracted Prompts
Prompts/LLM Interaction Classification.md#Prompt 1