NCA-GENL: Generative AI and LLMsDemo

Question

Which prompt-engineering technique is most effective at improving an LLM's performance on multi-step reasoning tasks?

A Few-shot prompting with unrelated examples.
B Retrieval-augmented generation without context.
C Chain-of-thought prompting with explicit intermediate steps.
D Zero-shot prompting with detailed task descriptions.

Explanation

Chain-of-thought prompting explicitly elicits intermediate reasoning steps, which helps an LLM structure and complete multi-step problems more reliably.

NCA-GENL: Generative AI and LLMsDemo

AstroTutor

Notes

Readiness

QuestionQ53

Community Discussion