Posted on

Filtered Vector Search in MyScale

Vector databases offer lightning-fast retrieval on similar objects stored in between billions of records. However, you may also be interested in searching for related objects that match a specific set of conditions, known as filtered vector search. With help from MyScale(opens new window), you can boost your filtered vector searches to a new level.

Most vector indexes or vector stores work as dedicated index services. They support a partial filtered vector search implementation of MongoDB query and projection operators (opens a new window where you can input a dictionary of conditions.

Posted on

Bring Your Knowledge Base Into OpenAI’s GPTs With MyScale

On November 6, 2023, OpenAI announced the release of GPTs. On this no-code platform, as a professional (or hobbyist) developer, you can build customized GPTs or chatbots using your tools and prompts, effectively changing your interactions with OpenAI's GPT. Previous interactions mandated using dynamic prompting to retrieve responses from GPT with LangChain (opens new window)or LlamaIndex (opens window. Now, the OpenAI GPTs handle your dynamic prompting by calling external APIs or tools.

This also changes how we (at MyScale) build RAG systems, from building prompts with server-side contexts to injecting these contexts into the GPTs model.

Posted on

Discover the Performance Gain With Retrieval Augmented Generation

Large Language Models (LLMs) are smart enough to understand context. They can answer questions, leveraging their vast training data to provide coherent and contextually relevant responses, no matter whether the topic is astronomy, history, or even physics. However, due to their inability to connect the dots and remember all the details, LLMs, especially the smaller models like llama2-13b-chat, can hallucinate even when the requested knowledge is in the training data.

A new technique, Retrieval Augmented Generation (RAG), fills the knowledge gaps, reducing hallucinations by augmenting prompts with external data. Combined with a vector database (like MyScale (opens new window)), it substantially increases the performance gain in extractive question-answering systems, even with exhaustive knowledge bases like Wikipedia in the training set.

To this end, this article focuses on determining the performance gain with RAG on the widely-used MMLU dataset. We find that both the performance of commercial and open source LLMs can be significanlty improved when knowledge can be retrieved from Wikipedia using a vector database. More interestingly, this result is achieved even when Wikipedia is already in the training set of these models.

Posted on

Teach Your LLM to Always Answer With Facts Not Fiction

Large Language Models are advanced AI systems that can answer a wide range of questions. Although they provide informative responses on topics they know, they are not always accurate on unfamiliar topics. This phenomenon is known as hallucination.

What Is Hallucination?

Before we look at an example of an LLM hallucination, let's consider a definition of the term "hallucination" as described by Wikipedia.com(opens a new window:

Posted on

What To Expect From Retrievel-Augmented Generation and Self-Hosted LLMs

Retrieval-augmented generation (RAG) is an AI framework designed to augment an LLM by integrating it with information retrieved from an external knowledge base. Based on the increasing focus RAG has garnered lately, it is reasonable to conclude that RAG is now a prominent topic in the AI/NLP (Artificial Intelligence/Natural Language Processing) ecosystem. Therefore, let's jump in and discuss what to expect from RAG systems when paired with self-hosted LLMs.

In the blog post titled: “Discover the Performance Gain with Retrieval Augmented Generation,” we investigated how the number of retrieved documents can improve the quality of LLM answers. We also described how the vectorized LLM based on the MMLU dataset, stored in a vector database such as MyScale, generates more accurate responses when integrated with contextually relevant knowledge and without fine-tuning the dataset.