Foundation, Instruction-Tuned and Reasoning LLMs CT-GENAI Practice Questions (Page 3)

Questions 11–12

1. 11application · medium

   A QA team wants to use an LLM to convert natural language bug descriptions into structured test cases. They have a dataset of 500 example pairs (bug description → test case). The team wants the model to learn the mapping and generalize to new descriptions. What is the most effective approach?

   • AUse a foundation LLM with a few-shot prompt containing 3-5 examples.
   • BUse a foundation LLM and provide all 500 examples in the prompt each time.
   • CUse a reasoning LLM and ask it to reason about each bug description before generating the test case.
   • DFine-tune an instruction-tuned LLM on the 500 example pairs.
   Reveal Solution Discussion Select an answer first
2. 12foundation · easy

   Which LLM type is most appropriate for a task that requires solving a multi-step logic puzzle?

   • AInstruction-tuned LLM
   • BReasoning LLM
   • CFoundation LLM
   • DAny LLM type would perform equally well
   Reveal Solution Discussion Select an answer first