Llm Advisor

What it does

The llm-advisor skill provides guidance and recommendations when working with Large Language Models (LLMs). It helps users refine prompts, troubleshoot issues, and understand LLM behavior. Specifically, it can analyze existing prompts for clarity and effectiveness, suggest alternative phrasing to improve results, and explain why an LLM might be producing unexpected output.

When to use it

• Prompt Refinement: You're not getting the desired response from your LLM prompt and need help improving its structure or content.
• Troubleshooting Unexpected Output: The LLM is generating responses that are nonsensical, irrelevant, or factually incorrect, and you want to understand why.
• Understanding LLM Behavior: You're curious about how different phrasing or instructions impact the LLM’s output and want insights into its decision-making process.
• Learning Prompt Engineering Techniques: You are new to prompt engineering and would like assistance in crafting more effective prompts.

Key capabilities

• Prompt analysis for clarity and effectiveness
• Suggestion of alternative phrasing for improved results
• Explanation of LLM behavior and potential causes of unexpected output

Example prompts

• "Analyze this prompt: 'Write a short story about a cat.' How can I make it more specific?"
• "Why did the LLM generate an incorrect answer when I asked, 'What is 2 + 2?'"
• "Suggest improvements to this prompt: 'Summarize this article.'"

Tips & gotchas

The llm-advisor skill works best with clear and concise prompts. While it can analyze complex prompts, simpler examples often yield more actionable feedback.