Input Optimization

What are Inputs?

Inputs are pieces of information that go into a prompt or model. There are 3 types of inputs:

1. Uploaded Inputs

2. End User Inputs

3. Inputs from your RAG

Uploaded Inputs

Uploaded inputs are pieces of information that you want in your prompt or model from your organization or team. Uploaded inputs can be:

1. Text

2. PDFs

3. Images

4. Audio

5. with more types coming in the future

if you don't see a type here please reach out to [e-mail]

End User Inputs

End User inputs are inputs that come from your end users or subject matter experts. If you use a:

1. Thumbs up or thumbs down user verification in your product

2. Have a human in the loop flow

Thread key

Rag Optimization

Optimize Prompts that drive the Rag system

recombines the rag documents with the rag system.

choose a standard eval - hallucination

replace rag prompt with prompt registry

optimize the hyper parameters of the rag

returning it in real-time and select the ideal subset of the question

For example how do I get my chunks

Input Anomalies

input anomaly score when you put in inputs automatically

sort all data an look at examples unexpected because it might be to diagnose unexpected behavior