A "Large Language Model," or LLM, is a type of artificial intelligence that's been trained on massive amounts of text data (we're talking billions of words here!) to generate human-like content based on the input they receive. LLMs use complex algorithms to analyze the patterns and structure of language, enabling them to generate responses that are contextually relevant and linguistically coherent. Essentially, this allows them to understand and respond to natural language inputs in a way that's similar to how a person might communicate.

Current LLMs (such as ChatGPT or Jasper) would be considered to be horizontally trained, enabling them to offer general knowledge across a broad range of topics. This is unlike vertically trained LLMs, which is a type of model that's been trained on data from a specific domain or field. It's like taking a general-purpose LLM and giving it a crash course in a particular subject. For example, you could have a medical LLM that's been trained on medical texts and data, or a legal LLM that's been trained on legal documents and statutes. This allows the LLM to generate specialized responses that are more accurate and tailored to the specific subject matter.

Vertically trained LLMs create an opportunity for Enterprises to combine their internal data assets with 3rd party datasets to create a more specific, holistic view of an industry, which in turn can identify competitive advantages. With a vertically trained LLM, industry specific acronyms will be taken into account and the training will be rapid compared to horizontally trained LLMs.

As we look forward, pre-built reports and analytics will remain mission critical, but vertical LLMs will become a complementary business intelligence tool that uncovers a wealth of insights. In fact, at Data Buddies, we have already had requests for our data pipeline solution to feed clean and harmonized data to vertical LLMs.

The future looks promising for vertical LLMs! As data sets in specific domains continue to grow, the need for specialized LLMs will increase. As companies and organizations continue to recognize the benefits of these specialized models, we predict that they will become increasingly popular and widely used.