This demo showcases how we can significantly improve an LLM's decision-making capabilities by introducing structured mathematical analysis into the prompt engineering process.

Try opening ChatGPT and playing Tic-Tac-Toe - you'll notice it often makes basic mistakes and struggles with game state tracking. This is because LLMs inherently have difficulty with precise mathematical reasoning and state management.

My solution implements my propriety Value Added Prompt Framework to enhance the LLM's capabilities by:

• Calculating precise mathematical scores for each possible move
• Translating these scores into natural language analysis
• Providing structured context for the LLM to make informed decisions