11: large language models

not the same but similar

Large language models and chatbots are?

chatbots

By fine-tuning LLM we can get?

GPT

Which is the prototypical LLM based chatbot?

175B

How many parameters does GPT-3 have?

Open source LLM’s

Llama and MistalAI är examples of?

hallucination, reasoning abilities, lack of robustness

What are three limitations of chatbots?

scale in approximately equal porportions

According to the Chinchilla Scaling Law, given an increased budget limit in flops, the number of parameters and number of tokens for training should?

low bit quantization

In order to fine tune LLM we need to fit it on GPU, what can we do?

LoRA

For efficient fine tuning we can use?

RAG models

extended chatbots and Q&A systems with capacity to:

• retrieve content in real time

• and generate responses based on retrieved content