Google

To enhance the clarity and reduce ambiguity of prompts, specific formatting is often applied. This practice helps in structuring the information for the LLM in a more readable way. For instance, a prompt can be structured with distinct fields for different types of information. Another common technique is to use a code-like syntax, especially for LLMs capable of processing both natural language and code.

Formatting Prompts for Clarity

To improve the clarity and readability of prompts while reducing ambiguity, it is a common practice to format them with distinct fields. These fields are identified by labels (e.g., `SYSTEM`, `USER`, `Input:`, `Output:`) that structure the prompt and clarify the role of each piece of information for the language model.

Prompt Field Labeling for Clarity and Structure

A code-like prompt template enhances clarity by structuring instructions and input in a format that resembles a programming language. This approach is particularly effective for Large Language Models capable of understanding and generating both natural language and code, and it serves as an alternative to other formats like question-answer or direct instruction templates.

Code-like Prompt Templates

The following prompt is poorly structured, making it potentially confusing for a language model. Rewrite the prompt to improve its clarity and reduce ambiguity by organizing the information into distinct, clearly defined sections.

**Original Prompt:**
'Write a product description for a new coffee mug. It's a 12-ounce ceramic mug, it's dishwasher safe, and the target audience is busy professionals. Make it sound witty and engaging, and it should be about 50 words.'

Refining a Prompt for Clarity

A developer has created the following prompt for a customer service chatbot. The goal is for the chatbot to act as a friendly and helpful assistant for 'InnovateTech'. It should use the provided conversation history to answer the user's latest question about their order status. 

Evaluate the current structure of this prompt. Identify at least two specific weaknesses in its formatting and explain how these weaknesses could lead to an unclear or incorrect response from the language model.

Evaluating Prompt Structure for a Chatbot

An engineer is tasked with creating a prompt that instructs a language model to summarize a given text. Below are two versions of the prompt they are considering.

**Prompt A:**
"Please summarize the following text for me. The summary should be exactly two sentences long and maintain a formal tone. The text to summarize is: The quick brown fox jumps over the lazy dog."

**Prompt B:**
"**INSTRUCTIONS:**
Summarize the provided text.

**CONSTRAINTS:**
- Length: Exactly 2 sentences.
- Tone: Formal.


Learn Before

Related