Please add Chain-of-Thought, RAGAS, any additional explainability approach
From Pranay's email with subject: Customer complaint analyser / Mortgage calculator
Use folllowing trusted AI approach
Show example of hallucination
Show how to solve it using embedding: Create embeddings of correct SQLs. Compare LLM Query with nearest actual query with embeddings
Show how to solve it using fine-tuning: Create a fine-tuned LLM which has list of predefined questions, and sql. The chatbot used the fine-tuned LLM instead of generic LLM
From Pranay's email:
From Pranay's email with subject: Customer complaint analyser / Mortgage calculator
Use folllowing trusted AI approach
PR# https://github.com/Teradata/jupyter-demos/pull/619