logseq/pages/Omnivore___01-11-2023___compact-guide-to-large-language-models.md

• full-title:: compact-guide-to-large-language-models site:: databricks.com labels:: LLMs date-saved:: 01-11-2023 date-archived:: 17-11-2023 is-archived:: 10 source:: Omnivore state:: archived

  • • Highlights

    collapsed:: true
    • • id:: d8773d18-2617-49a9-aa6a-1efd7babdf52

      As the world

	  becomes more data driven, everything, including LLMs, will be reliant on having 

	  a strong foundation of data. [⤴️](https://omnivore.app/me/compact-guide-to-large-language-models-18b8b828feb#d8773d18-2617-49a9-aa6a-1efd7babdf52)
	- * id:: c74610d9-4112-44a8-b26a-46f0829e80dc
	  > MLflow to make it very easy for someone with a bit of Python 

	  experience to pull any Hugging Face transformer model and use it as a Python 

	  object [⤴️](https://omnivore.app/me/compact-guide-to-large-language-models-18b8b828feb#c74610d9-4112-44a8-b26a-46f0829e80dc)
	- * id:: 95b9223f-a43a-4ac2-9167-398d7e071a97
	  > Large language models are AI systems that are designed to process and analyze vast amounts of natural language data and then use that information to generate  responses to user prompts. [⤴️](https://omnivore.app/me/compact-guide-to-large-language-models-18b8b828feb#95b9223f-a43a-4ac2-9167-398d7e071a97)
	- * id:: 82aee9cc-a844-45de-933f-7335bff79899
	  > This means that you can 

	  keep the data in your control for privacy and governance concerns as well as 

	  manage your costs. [⤴️](https://omnivore.app/me/compact-guide-to-large-language-models-18b8b828feb#82aee9cc-a844-45de-933f-7335bff79899)
	- * id:: 386602a1-9c25-40d8-b8b3-bbb95f958b8a
	  > While this works sometimes, strictly defined rules only work for  concrete, well-defined tasks that the system has knowledge about. [⤴️](https://omnivore.app/me/compact-guide-to-large-language-models-18b8b828feb#386602a1-9c25-40d8-b8b3-bbb95f958b8a)
	- * id:: 6a819177-5d44-45c2-9334-72000f327ec7
	  > learn the patterns and structures  of human language, and are capable of generating natural language responses to  a wide range of written inputs [⤴️](https://omnivore.app/me/compact-guide-to-large-language-models-18b8b828feb#6a819177-5d44-45c2-9334-72000f327ec7)
	- * id:: bbdaeaa4-2d5f-4ffe-b9d6-4d2719da4e1a
	  > OpenAI releases GPT-3, which becomes the largest model at  175B parameters [⤴️](https://omnivore.app/me/compact-guide-to-large-language-models-18b8b828feb#bbdaeaa4-2d5f-4ffe-b9d6-4d2719da4e1a)
	- * id:: 7614d4e0-6e6c-4970-bff8-351fb376095e
	  > statistical models and  language patterns [⤴️](https://omnivore.app/me/compact-guide-to-large-language-models-18b8b828feb#7614d4e0-6e6c-4970-bff8-351fb376095e)
	- * id:: 77ab6ce2-ed73-45b5-a73f-b468497254d7
	  > there are techniques that let you take open source models and train 

	  them to your specific data, greatly improving their performance on your 

	  specific domain. [⤴️](https://omnivore.app/me/compact-guide-to-large-language-models-18b8b828feb#77ab6ce2-ed73-45b5-a73f-b468497254d7)
	- * id:: 2323f710-a579-403a-8863-c32362ae0382
	  > one aspect under a larger generative AI umbrella. [⤴️](https://omnivore.app/me/compact-guide-to-large-language-models-18b8b828feb#2323f710-a579-403a-8863-c32362ae0382)
	- * id:: fcf3a571-b396-452c-a7da-a3cb17f98b7f
	  > Most LLMs are not trained to be fact machines. They know how to use  language, but they might not know who won the big sporting event last year.  It’s always important to fact check and understand the responses before  using them as a reference. [⤴️](https://omnivore.app/me/compact-guide-to-large-language-models-18b8b828feb#fcf3a571-b396-452c-a7da-a3cb17f98b7f)
	- * id:: 0eabe4a8-4f18-4286-b37d-26e6c8516e45
	  > e LLMs begin showing increasingly impressive results with releases such as Dolly 2.0, LLaMA, Alpaca and Vicuna. [⤴️](https://omnivore.app/me/compact-guide-to-large-language-models-18b8b828feb#0eabe4a8-4f18-4286-b37d-26e6c8516e45)
	- * id:: ff58d2db-751f-444a-8e5e-68e5a08f8500
	  > largest jumps in performance has come from integrating  human feedback directly into the training process [⤴️](https://omnivore.app/me/compact-guide-to-large-language-models-18b8b828feb#ff58d2db-751f-444a-8e5e-68e5a08f8500)
	- * id:: 256ad1d9-b4d2-4386-92be-f427c6ef93ed
	  > (Generative Pre-trained Transformer) [⤴️](https://omnivore.app/me/compact-guide-to-large-language-models-18b8b828feb#256ad1d9-b4d2-4386-92be-f427c6ef93ed)