Explore the inner workings of ChatGPT and LLMs, and discover three effective approaches for building generative AI search experiences for specific domains.

What is ChatGPT?

Undestanding Large Language Models (LLMs)

LLMs are expensive to train

How does a GPT model work?

Tokens

Embedding

Transformer

The challenges with ChatGPT and LLMs

Integration of Elasticsearch with LLMs

Method 1: Provide context to your data and integrate with ChatGPT or other LLMs

Method 2: Bring your own model

Method 3: Use the built-in sparse encoder model

Terms and definitions

Neural Network (NN)

LLM parameters

SPLADE

Wrap up

The GPT model uses a technique called “attention” to focus on the most relevant parts of the input and output texts and to capture long-range dependencies and relationships between words. The model is trained by using a large corpus of texts as both the input and the output and by minimizing the difference between the predicted words and the actual words. It can then be fine-tuned or adapted to specific tasks or domains by using smaller and more specialized data sets.

Elasticsearch can be used as a vector database, and to perform hybrid retrieval across text and vector data. There are three patterns where Elasticsearch can provide clear benefits when used with LLMs: 1) Provide context to your data and integrate with ChatGPT or other LLMs. 2) Enable you to bring your own model (any 3rd party model). 3) Use the built-in Elastic Learned Sparse Encoder model.

How does Elasticsearch fit with LLMs?

Demystifying ChatGPT & LLMs: Methods for building AI search experiences

Demystifying ChatGPT & LLMs: Different methods for building AI search

Learn how to implement and scale semantic search as a service for a search Center of Excellence (COE) using ELSER.

Author

Sherry Ger

Articles

Demystifying ChatGPT & LLMs: Different methods for building AI search

Semantic search as service at a search center of excellence