The Benefits of Retrieval Augmented Generation

I’ve been working on AI the Docs for a few weeks and I just opened it up for Early Adopters this week. The secret behind the app is Retrieval Augmented Generation (RAG). With RAG, you add context to the prompt that gives an LLM more information and allows it to give a more accurate answer.

During development, an LLM is trained on a bunch of content. At some point, the training has to be cut off and it doesn’t know anything new, but RAG lets you tell the LLM about things that it doesn’t already know.

How about an example…

The Setup

Laravel Pulse was released in December 2023. I asked both ChatGPT and AI the Docs about Laravel Pulse.

With ChatGPT 3.5

As you can see, it doesn’t know anything about Laravel Pulse. Its training data isn’t nearly recent enough and it has no way to get additional information.

With AI the Docs

Since AI the Docs knows about all the Laravel 10 documentation, it can include relevant information related to Pulse in the prompt for a much more helpful answer.

Wrap Up

Even though under the hood both chat interactions used the ChatGPT 3.5 client, AI the Docs gives a much better answer because it is able to leverage RAG to provide additional context to the LLM for a better, more accurate answer.

You can sign up as an Early Adopter and try AI the Docs today!