LLM Primer Part 1
Please see the source reference for the information in this blog. This is a high-level summary of a fuller summary at HTTP://www.databricks.com.
The following is a guide to understanding the development history of LLMs, which is the basis for Artificial Intelligence generated content.
Definition: LLMs (large language models) are AI systems that are designed to process and analyze natural language data and then use that to generate prose for the "prompt" that it has been given. The prompt is basically a search phrase, and the better the prompt --- the better the results. There truly is an art to creating the right prompt.
The system algorithms for this are focused on patterns and structures of language that are capable of generating natural language in response to the prompts.
LLM is not a recent development . . .
1950s to 1990s
Initial attempts are made to map hard rules around various languages . . .
1990s
Language models begin evolving into stat models and language patterns that start being analyzed . . .
2000s
Advanced machine learning increases the complexity of language models . . .
2012
Advances in deep learning architecture and larger data sets that lead to developing GPT (Generative Pre-trained Transformer)
2018
Google introduces BERT (Biderectional Encoder Representations from Transformers), which paved the way for the future.
2000s
Advancements in machine learning (1) increases the complexity of large language models.
2012
Advancements occurred in deep learning architectures and larger data sets lead to the development of GPT (generative pre-trained transformer). (See my previous blogs about Wikipedia use for this)
2018
Google introduces BERT (bidirectional encoder representatives from Transformers) that is the future of LLMs.
2020
Open AI releases GPT-3, which becomes the largest model and sets a new performance or benchmark for LLM tasks.
2022
ChatGPT is launched and becomes widely accessible to users through a web interface and creates a huge awareness of AI, LLMs, and chatbots.
2023
Open source (2) LLMs begin showing impressive results and version GPT-4 is also released, setting new benchmarks for size and performance.
This blog source is a summary of "A Compact Guide to Large Language Models. (3)
# # # #
Comments