Demo Speaker Notes

pipeline intro

1. so discuss the overview of why each step is needed.
2. why we choose these specific steps.

System prompt generator

1. This is how we inject custom information for a usecase to the LLM.

2. Talk about view and update :- user control.

3. Testing the system prompt.

Information extraction

1. We only show the extraction rule. why? what is the extraction rule used for?

2. a comment on the fact that we can augment the testing rule like the previou step.

3. show that the pipeline is step by step.

SPARQL query generator.

1. we can define the end point and the query to retrieve the information.
2. here we have have a place where a single item is injected to get the information

Post - processing

1. we sub divide this part to two main parts one processe\

One instruction to post processes thw sqparql query , duplication removal - NA handling formatting:- result aggregation and column names

pipeline building.