Generative AI and Large Language Models (LLMs) AIE Practice Questions (Page 5)

Questions 21–25

1. 21foundation · easy

   Which of the following is a common capability of large language models?

   • AGuaranteeing factual accuracy in all outputs
   • BExecuting operating system commands
   • CSummarizing long documents into concise summaries
   • DPerforming high-precision arithmetic calculations
   Reveal Solution Discussion Select an answer first
2. 22foundation · easy

   Which of the following tasks best exemplifies generative AI?

   • AGenerating a realistic image of a cat that does not exist
   • BIdentifying whether an email is spam or not spam
   • CTranslating a sentence from English to French
   • DDetecting objects in a photograph
   Reveal Solution Discussion Select an answer first
4. 23foundation · easy

   How does fine-tuning differ from pre-training in the context of LLMs?

   • AFine-tuning adapts a pre-trained model to a specific task or domain
   • BFine-tuning removes the need for any training data
   • CFine-tuning changes the model's architecture to a different type
   • DFine-tuning uses a much larger dataset than pre-training
   Reveal Solution Discussion Select an answer first
5. 24foundation · easy

   What is the primary goal of generative AI compared with discriminative AI?

   • ATo create new data instances that resemble the training data
   • BTo compress data for efficient storage and transmission
   • CTo predict the next value in a time series
   • DTo classify input data into predefined categories
   Reveal Solution Discussion Select an answer first
6. 25foundation · easy

   What is the purpose of self-attention in the transformer architecture?

   • ATo convert tokens into dense vector representations
   • BTo encode the position of each token in the sequence
   • CTo reduce the vocabulary size of the model
   • DTo allow each token to weigh the importance of other tokens in the sequence
   Reveal Solution Discussion Select an answer first