AI -- LLMs summary - March 2025
Last updated
Was this helpful?
Last updated
Was this helpful?
AI -- LLMs summary - March 2025
LLMs (Large Language Models) are advanced AI models designed to understand, analyze, process and generate natural human language. They are trained on massive amounts datasets and use deep learning techniques, such as transformer architectures, latest mixture of expert (MoE) models, MAMBA, etc to perform a wide range of tasks such as text generation, summarization, translation, and more.
(As a part-time AI professor in some Asia universities, I teach such AI models, AI policies, AI trends, AI digital transformation consulting strategies, AI business, AI for good, AI data security, etc.)
Types of LLMs:
Domain-Specific LLMs – Some models specialize in legal, medical, or financial sectors, like BloombergGPT for finance.
General-Purpose LLMs – Models like Llama, GPT-4, Gemini, and Claude are designed for a broad range of applications.
Open-Source LLMs – DeepSeek, Meta’s LLaMA, Mistral, and Falcon provide publicly available models.
Efficient & Small Language Models – Models like Phi-3 and Gemma are optimized for low-resource environments.