What Is ChatGPT Doing... and Why Does It Work?(2023.2)

Resource link:

It's just adding one word at a time
- the newly generated word is input as context to the model again to generate the next word.
Where Do the Probabilities Come From? from the llm
What is a model? It's a function with billions of parameters, the function and parameters are learned from the huge amount of data.

Last updated 2 years ago