Large Language Models: A short introduction to the core concepts of transformers and LLM training

0:00 Introduction: The different types of LLMs and overview of the presentation.
0:33 Transformers: Deep learning, BERT, GPT, word embedding, attention, and text generation.
2:03 Training base models: Self-supervised learning, masked training, and how LLMs learn facts.
3:17 Fine tuning: Supervised learning, manually annotated corpora, named entity recognition, and information extraction.
4:38 Instruction tuning: Chat templates, the broad scope of instruct models, and their downsides.
6:01 Summary: LLMs, transformers, base models, fine-tuned models, and instruct models.

[zipfslaw]

Thank you from Ukraine, where you give us a nice break from thinking about ruzzian war crimes!