Ive recently created a Text2SQL Application that allows Retail Executives to ask natural language questions about their business performance metrics housed in a Snowflake database. In addition to performing Text2SQL LLM Prompting, I’ve used some other Meta Prompting Techniques that I would like to introduce to the audience.
- LLM as a Prompt Optimizer
- Iterative, self correcting prompting by an LLM
- LLM as a judge
All of these techniques are founded in prior research within the domain.
