LLM Data Process

A large language model works by ______ what word comes next in a sequence.

• answering

• predicting

• asking

• telling

How do models learn how to make these predictions?

• They process information pulled from the internet.

• They use random data.

• They just guess with no knowledge.

• They ask people.

The continuous values that determine systems' behaviors used are called...

• choices

• values

• parameters

• data

What is backpropagation used for in the training of language models?

• To speed up the training process

• To refine the model’s parameters

• To generate new text

• To evaluate the model's performance

What does RLHF stand for?

Transformers read text in ___

Each word has a corresponding ___ that is encoded in the system.

• language

• number

• word

• form

What are the 2 steps that are iterated to accurately predict words?

• Attention

• Refining

• Feed-forward

• Trial & Error