LLMs Foundations: Tokenization and Word Embeddings Models