What documents can I upload?

The more complex the file, the more processing power it requires. To be more economical, we suggest pre-processing files and saving their text content to a “.txt” file. For instance, if you have a PDF or Word file that only contains text, open it, copy all txt ([CTRL] + a in Windows or [CMD] + a in Mac) Then, create a text (e.g., .txt or .rtf) file. [TODO: add instructions] – Text – Markdown – PDF – Powerpoint – CSV – Word – Audio – Video You can also create API-based brains that will draw data from another app’s API.

What does temperature mean?

Imagine you’re telling a story with a friend who helps decide what happens next. The "temperature" setting controls how *wild or predictable* your friend’s ideas are. Low Temperature (like 0.1): Your friend only suggests things that are super obvious or safe, like “then the hero wins easily.” It’s boring but makes total sense—no surprises! This is the temperature you would use for scientific data or any kind of research that should be only fact-based and unbiased. Use Case: This is a good choice when dry, clinical responses are required. for instance when doing legal research or when analyzing financial or medical data. Medium Temperature (like 0.7): Your friend balances ideas—they might say, “the hero uses a cool trick to win!” It’s creative but still logical. This is what most people use because it’s just right. Use Case: Most people use this for emails or ideas—they want something smart but not too crazy. High Temperature (like 1.5+): Your friend starts getting *weirdly creative*—they might suggest the hero fights a giant marshmallow or suddenly turns into a cat. The story becomes fun but might not make much sense anymore. Use Case: Artists or game designers might use this for fun, weird ideas—but then they’ll clean up the nonsense later. Think of temperature like a volume knob for creativity: turn it down for safety, crank it up for crazy ideas (but expect some nonsense). Use this to decide how “silly” or “smart” an AI should act. Temperature Range: The typical temperature range for AI models is between 0 and 2 . Minimum: Technically , it can’t be 0 because of calculation issues (dividing by zero). In practice, it’s often set to a very low value like 0.1 or slightly above , which makes the model pick the most likely choice almost every time. Maximum: The upper limit is usually 2 , beyond which outputs may become too random or nonsensical. For most users, setting temperatures between 0.1 and 2 works well: 0.1 : Predictable & safe 0.7 : Balanced (creative but logical) 1 : Default “standard” randomness 2 : Very creative (or chaotic)

What is an LLM (Large Language Model)?

A Large Language Model (LLM) is a powerful type of artificial intelligence system designed to understand and generate human language. It’s made up of a vast number of interconnected virtual “neurons” that can process and generate text. LLMs like GPT-3, for example, can handle a wide range of natural language tasks, such as answering questions, translating languages, writing articles, and even simulating human-like conversations. These models are trained on massive datasets, allowing them to learn patterns and nuances in language, making them valuable tools for various applications in fields like natural language processing, machine learning, and text generation.

What is custom language model training?

Most public language models are trained on vast amounts of data, which is “scraped” from all the data that is fed into the model. Every time you send something to OpenAI’s ChatGPT, for instance, OpenAI uses your data to “train” its model. The intent is to provide a broader knowledge base, which may be more useful for most common interactions with ChatGPT. For some uses, a more specific language model may be needed, which is trained on a narrower set of data. E.g., a law office may choose to train its own language model on the kind of data it primarily works with. This will improve the responses that come from the language model, since it reduces the risk of getting unrelated information. Yes, even language models can get confused, or don’t understand the specific context of a question. Training your own language model can help.

What is the difference between a local language model and OpenAI?

The terms “local language model” and “OpenAI” refer to different aspects of language models, so let’s clarify their meanings: Local Language Model: This term typically refers to a language model that operates on a local or on-premises system, meaning it runs on your own computer or server, rather than relying on external cloud-based services. Local language models can be fine-tuned and customized to suit specific needs or security requirements, and they might be used for various tasks, such as text generation, translation, or chatbots. They are often used when data privacy and control are paramount. OpenAI: OpenAI is an artificial intelligence research organization that has developed various language models, including GPT-3 (and possibly newer versions as of my last knowledge update in January 2022). OpenAI’s models are known for their general language understanding and text generation capabilities. OpenAI provides access to its models through APIs, allowing developers to integrate their models into applications and services. So, the key difference is that “OpenAI” refers to the organization developing the language models, while “local language model” refers to where and how the model is deployed and used. You can use an OpenAI model both locally and through cloud-based services, depending on your needs and preferences.

AI agents are self-contained bodies of information. They can be used to provide context to Large Language Models (LLMs) and answer questions on a particular topic. LLMs are trained on a large variety of data. However, to answer a question on a specific topic, or to draw conclusions from a specific topic, they need to be supplied with the context of that topic. AI agents are an intuitive way to provide that context. Selecting a agent provides the context of that agent to the LLM. This lets users build agents for specific topics, and then use them to answer questions about that topic, without being “polluted” by information that is not in the agent. This prevents “hallucinations” and answers that are out-of-context.

FAQs Archive - Modular Technology Group

What documents can I upload?
The more complex the file, the more processing power it [...]

Continue reading
What does temperature mean?
Imagine you’re telling a story with a friend who helps [...]

Continue reading
What is an LLM (Large Language Model)?
A Large Language Model (LLM) is a powerful type of [...]

Continue reading
What is custom language model training?
Most public language models are trained on vast amounts of [...]

Continue reading
What is the difference between a local language model and OpenAI?
The terms “local language model” and “OpenAI” refer to different [...]

Continue reading
What is an AI agent?
AI agents are self-contained bodies of information. They can be [...]

Continue reading

12 Next