Process / pipeline
提示工程 — 大型语言模型的指令设计
提示工程是精心设计结构化自然语言指令（即提示）以从大型语言模型（LLMs）获得目标输出的实践。该方法由 Brown 等人（2020）在 GPT-3 的背景下形式化，并由 Wei 等人（2022）通过思维链提示（chain-of-thought prompting）进行了扩展，它包含四种主要策略：零样本（zero-shot）、少样本（few-shot）、思维链（chain-of-thought）和思维树（tree-of-thought）。分析师不重新训练模型，而是完全通过输入文本的设计来塑造模型的行为。
阅读完整方法
仅限会员
登录使用免费账户登录即可阅读本节。
Method map
The neighbourhood of related methods — select a node to explore.
来源
- Brown, T. et al. (2020). Language Models are Few-Shot Learners. Advances in Neural Information Processing Systems (NeurIPS), 33, 1877-1901. link ↗
- Wei, J. et al. (2022). Chain-of-Thought Prompting Elicits Reasoning in Large Language Models. Advances in Neural Information Processing Systems (NeurIPS), 35. link ↗
如何引用本页
ScholarGate. (2026, June 1). Prompt Engineering (Instruction Design for Large Language Models). ScholarGate. https://scholargate.app/zh/text-mining/prompt-engineering
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- 少样本文本分类文本挖掘↔ compare
- GPT模型微调深度学习↔ compare
- LoRA 和 PEFT深度学习↔ compare
- 自然语言生成文本挖掘↔ compare
- 检索增强生成（RAG）文本挖掘↔ compare
- 文本分类文本挖掘↔ compare
- 零样本分类文本挖掘↔ compare