Home Library Explore SAT

Home Library Explore SAT

Summer School recap Flashcards | StudBud

Explore/Summer School recap

Summer School recap

7 terms·StudBud Administration·4 plays

Imported from external platform

Flashcards7 cardsFPractice quizPHost liveL

Questions 7

Answers

1
What is one advantage of scaling up LLMs?
Need for more training examples per task
Emergence of new, untrained capabilities
Less accurate prompt handling
Reduced generalization capability
2
Which of the following best describes an embedding?
A compressed version of the training data
A vector representation capturing the meaning of a word or sentence
A visual representation of a neural network
A programming language used in NLP
3
Why is prompting useful in large language models?
It speeds up data labeling
It allows humans to adjust training weights manually
It is used to clean datasets before training
It gives instructions that guide the model's behaviour on-the-fly
4
What is the main goal of fine-tuning a language model?
To train the model from scratch
To adapt the model to a specific task or domain
To create embeddings for unknown words
To compress the model for faster downloads
5
What is the best definition of an Artificial Intelligence system?
A system that mimics humans by learning, reasoning, making decisions
A tool that follows predefined instructions without adaptation
A type of hardware designed to improve computer graphics
A system that stores and retrieves large volumes of data
6
What is the primary function of a Language Model (LM)?
Understanding and generating human language
Simulating neural activity
Classifying images
Storing large text datasets
7
What makes Transformers efficient for training large models?
Use of handcrafted rules
Use of complex heuristics
Attention mechanism allowing parallel processing
Sequential processing