Generative AI and Large Language Models (LLMs) AIE Practice Questions (Page 7)

Questions 31–35

1. 31expert · hard

   A government agency wants to use an LLM to generate public information summaries. They are concerned about the model producing biased or misleading content. They have a strict requirement for transparency and accountability. Which approach is most aligned with these requirements?

   • AImplement a human-in-the-loop review process where all generated summaries are checked by a subject-matter expert before publication.
   • BOnly use the LLM for internal drafts and never publish its output.
   • CDeploy the LLM with no oversight to maximize efficiency.
   • DUse a larger LLM to ensure the content is unbiased.
   Reveal Solution Discussion Select an answer first
2. 32application · medium

   A researcher is comparing an LLM with a traditional n-gram language model. Which statement correctly highlights a key advantage of the LLM?

   • AThe LLM always produces grammatically correct sentences, while n-gram models do not.
   • BThe LLM can capture long-range dependencies in text through self-attention, while n-gram models are limited by fixed context windows.
   • CThe LLM does not require training data, while n-gram models do.
   • DThe LLM is faster to train than an n-gram model.
   Reveal Solution Discussion Select an answer first
4. 33application · medium

   A healthcare startup has a pre-trained LLM that performs well on general text but needs it to understand medical terminology and generate clinical notes in a specific format. They have a small dataset of de-identified clinical notes. What is the most appropriate approach?

   • AFine-tune the pre-trained LLM on the clinical notes dataset.
   • BUse a rule-based system to generate clinical notes.
   • CUse the pre-trained model without any adaptation and hope it generalizes.
   • DTrain a new LLM from scratch on the clinical notes dataset only.
   Reveal Solution Discussion Select an answer first
5. 34application · medium

   A developer is using an LLM to translate English technical documentation into Japanese. The output is grammatically correct but uses informal phrasing that is inappropriate for official documentation. Which prompt adjustment is most likely to improve the output?

   • AIncrease the temperature to 1.5 to encourage more varied vocabulary.
   • BUse a smaller model to reduce the chance of over-formalizing the output.
   • CAdd a style instruction such as 'Use formal, technical Japanese suitable for official documentation.'
   • DRemove the source text and ask the model to translate from memory.
   Reveal Solution Discussion Select an answer first
6. 35foundation · easy

   What is the primary purpose of pre-training a large language model?

   • ATo specialize the model for a specific task like sentiment analysis
   • BTo reduce the model's size for deployment on edge devices
   • CTo label a dataset for supervised learning
   • DTo learn general language patterns from a large, diverse text corpus
   Reveal Solution Discussion Select an answer first