We’ll also see a big surge in the use of buzzword-heavy AI concepts like Retrieval-Augmented Generation (RAG) systems, generative AI, and cloud-based AI products, all of which will become easier to use and, hopefully, cheaper, thereby driving further broad adoption.
RAG will shine even more in 2025
So how does this work? I wanted to give this picture of what's actually happening behind the scenes, especially with this question and answer. So first, I will say that we're using a combination of OpenAI's GPT 3.5 to do this as well as some open source, smaller open source models to generate the vectors for the semantic search.
This illustration shows four alternative ways to nudge an LLM to produce relevant responses:Generic LLM - Use an off-the-shelf model with a basic prompt. The results can be highly variable, as you can experience when e.g. asking ChatGPT about niche topics. This is not surprising, because the model hasn’t been exposed to relevant data besides the small prompt.Prompt engineering - Spend time structuring the prompt so that it packs more information about the desired topic, tone, and structure of the response. If you do this carefully, you can nudge the responses to be more relevant, but this can be quite tedious, and the amount of relevant data input to the model is limited.Instruction-tuned LLM - Continue training the model with your own data, as described in our previous article. You can expose the model to arbitrary amounts of query-response pairs that help steer the model to more relevant responses. A downside is that training requires a few hours of GPU computation, as well as a custom dataset.Fully custom LLM - train an LLM from scratch. In this case, the LLM can be exposed to only relevant data, so the responses can be arbitrarily relevant. However, training an LLM from scratch takes an enormous amount of compute power and a huge dataset, making this approach practically infeasible for most use cases today.
RAG with a generic LLM - Insert your dataset in a (vector) database, possibly updating it in real time. At the query time, augment the prompt with additional relevant context from the database, which exposes the model to a much larger amount of relevant data, hopefully nudging the model to give a much more relevant response. RAG with an instruction-tuned LLM - Instead of using a generic LLM as in the previous case, you can combine RAG with your custom fine-tuned model for improved relevancy.
Fine-tuning takes a pre-trained LLM and further trains the model on a smaller dataset, often with data not previously used to train the LLM, to improve the LLM’s performance for a particular task.
LLMs can be extended with both RAG and Fine-Tuning Fine-tuning is appropriate when you want to customize a LLM to perform well in a particular domain using private data. For example, you can fine-tune a LLM to become better at producing Python programs by further training the LLM on high-quality Python source code.
In contrast, you should use RAG when you are able to augment your LLM prompt with data that was not known to your LLM at the time of training, such as real-time data, personal (user) data, or context information useful for the prompt.
Vector databases are used to retrieve relevant documents using similarity search. Vector databases can be standalone or embedded with the LLM application (e.g., Chroma embedded vector database). When structured (tabular) data is needed, an operational data store, such as a feature store, is typically used. Popular vector databases and feature stores are Weaviate and Hopsworks that both provide time-unlimited free tiers.
RAG LLMs can outperform LLMs without retrieval by a large margin with much fewer parameters, and they can update their knowledge by replacing their retrieval corpora, and provide citations for users to easily verify and evaluate the predictions.