In this video, you will learn the concept of Embedding using a real life analogy:

– Recall Tokenization concept: 0.18 to 1.03 (reference video: https://youtu.be/fwUgGQHUvS0?si=7aRZRSCI_O_d3P09)
– Embedding concept through a real-life analogy: 1.30 to 2.35 (reference video https://youtu.be/vPZjrlbL8w8?si=0JP4XWZSOI5q_dOG)
– Visual representation: 2.38 to 2.58
– Analogy comparison with Tokens embedding: 3.00 to 4.00
– Attention mechanism role to bring contexual info in weight values: 4.02 to 5.15 (reference video https://youtu.be/vPZjrlbL8w8?si=0JP4XWZSOI5q_dOG)
– In general visual representation of any token: 6.30 to 7.50
– Summary: 7.55 to 8.12

For detailed content on this topic, refer my blog here

Embedding: The language of AI World (2/2)

Other useful references:

Attention mechanism series by serrano.academy

#embedding #tokenization #llm #ai #vector #largelanguagemodels #chatgpt #vector

source