1. 🔍 Parsing User Intent

The LLM is trained to recognize the intent behind user input, even when phrased in diverse, non-technical, or casual ways.

Examples:

User Input	Detected Intent
“Deploy a token called GoldCoin with 5M supply”	deploy_erc20
“What’s my balance of GDC?”	get_token_balance
“Transfer 100 DIAI to my friend”	transfer_tokens

The model is able to:

• Understand imperative commands (“Deploy this”, “Send that”)

• Handle indirect requests (“Can I get my MYT balance?”)

• Interpret incomplete prompts and ask for clarification