Automatic Prompt Engineer (APE)
Image Source: Zhou et al., (2022) (opens in a new tab)
Zhou et al., (2022) (opens in a new tab) propose automatic prompt engineer (APE) a framework for automatic instruction generation and selection. The instruction generation problem is framed as natural language synthesis addressed as a black-box optimization problem using LLMs to generate and search over candidate solutions.
The first step involves a large language model (as an inference model) that is given output demonstrations to generate instruction candidates for a task. These candidate solutions will guide the search procedure. The instructions are executed using a target model, and then the most appropriate instruction is selected based on computed evaluation scores.
APE discovers a better zero-shot CoT prompt than the human engineered "Let's think step by step" prompt (Kojima et al., 2022 (opens in a new tab)).
The prompt "Let's work this out in a step by step way to be sure we have the right answer." elicits chain-of-thought reasoning and improves performance on the MultiArith and GSM8K benchmarks:
Image Source: Zhou et al., (2022) (opens in a new tab)
This paper touches on an important topic related to prompt engineering which is the idea of automatically optimizing prompts. While we don't go deep into this topic in this guide, here are a few key papers if you are interested in the topic:
- Prompt-OIRL (opens in a new tab) - proposes to use offline inverse reinforcement learning to generate query-dependent prompts.
- OPRO (opens in a new tab) - introduces the idea of using LLMs to optimize prompts: let LLMs "Take a deep breath" improves the performance on math problems.
- AutoPrompt (opens in a new tab) - proposes an approach to automatically create prompts for a diverse set of tasks based on gradient-guided search.
- Prefix Tuning (opens in a new tab) - a lightweight alternative to fine-tuning that prepends a trainable continuous prefix for NLG tasks.
- Prompt Tuning (opens in a new tab) - proposes a mechanism for learning soft prompts through backpropagation.