Boost Your AI Search Skills: Advanced Prompt & Query Tricks

Introduction

This article introduces how to use AI tools, especially large models with deep web‑search abilities, to enhance the efficiency and accuracy of information retrieval.

1. Site‑specific Search (site Syntax)

The site: operator restricts results to a particular domain, e.g., site:huawei.com/cn to find the latest Huawei products from the official site. Similar usage applies to government domains using .gov.cn.

2. Wildcard Usage

Wildcards help refine queries when only partial information is known.

+ : Both keywords must appear. Example: 西双版纳+大象 (search for both terms).

- : Exclude a keyword. Example: 旅游 -海滩 (exclude “beach”).

? : Matches a single character. Example: 张家? (matches “张家口” or “张家界”).

* : Matches any number of characters. Example: *ultra (matches any model ending with “ultra”).

3. Date‑Range Limitation

Use datarange: or .. to restrict results to a specific time span, e.g., searching for Zhejiang housing prices from 2023 to 2025.

4. Boolean Logic

Combine conditions with logical operators.

AND (与): Both terms must be present. Example: 广州 AND 火锅.

OR (或): Either term may appear. Example: 小米su7 OR 小米su7 ultra.

NOT (非): Exclude a term. Example: su7 NOT 小米 su7 ultra.

Prompt Techniques

The article summarizes key prompt engineering concepts from Su Jiang’s “Learn to Ask – Key Skills for AI Large‑Model Dialogue”. It covers Socratic questioning, Warren Berger’s questioning methods, and a structured prompt framework (Instruction, Context, Input Data, Output Indicator). Advanced CRISPE framework components are also described:

C – Capacity and Role: Define the AI’s role and abilities.

I – Insight: Provide background and specific user needs.

S – Statement: State the task clearly.

P – Personality: Specify the desired tone or style.

E – Experiment: Ask for multiple solutions or alternatives.

References

[1] https://docs.feishu.cn/v/wiki/EWl7wxKSQiLtndk2rLNceIr1nQg/aa [2] Su Jiang, “Learn to Ask – AI Large‑Model Era and ChatGPT Dialogue Key Skills” [3] Wei, J., Wang, X., Schuurmans, D., et al. (2022). Chain of Thought Prompting Elicits Reasoning in Large Language Models.