15,049
edits
AI Prompt Engineering: Difference between revisions
Jump to navigation
Jump to search
→Meeting Minutes
| (12 intermediate revisions by the same user not shown) | |||
| Line 1: | Line 1: | ||
AI Prompt Engineering: A Practical Guide | |||
{{Template:Generative AI Tool}} | {{Template:Generative AI Tool}} | ||
{{LanguageSwitcher | content = [[ | {{LanguageSwitcher | content = [[AI Prompt Engineering | English]], [[AI Prompt Engineering in Mandarin | 漢字]] }} | ||
== Automatic Prompt Generation and Enhancement == | |||
Use AI platforms to automatically generate and optimize prompts based on your requirements: | |||
* {{Gd}} Anthropic [https://platform.claude.com/dashboard Claude Developer Platform] | |||
# [https:// | *# How to: Select 'Generate a prompt' to optimize for Claude reasoning or non-reasoning models | ||
*# Pricing: The Console/API pricing is completely separate from Claude.ai subscriptions (Pro, Max, Team). When using "Generate a prompt," generating a structured prompt costs only a few cents, and testing a prompt costs a fraction of a cent per test. Additionally, the free tier provides roughly $10 USD in monthly credits, so light usage may not cost anything extra at all.<ref>[https://nickgarnett.substack.com/p/the-anthropic-console-a-practical (1) The Anthropic Console: A Practical Review for Prompt Writers]</ref> | |||
* [https://platform.openai.com/chat/edit?models=gpt-5&optimize=true OpenAI API - Edit Prompt] - Optimize prompts for OpenAI o3, GPT-4.1, GPT-5, GPT-5.1, and GPT-5.2 models | |||
* Google [https://docs.cloud.google.com/vertex-ai/generative-ai/docs/learn/prompts/prompt-optimizer Vertex AI prompt optimizer] | |||
== | == Converting Conversations into Reusable Prompts == | ||
<pre> | <pre> | ||
| Line 36: | Line 40: | ||
2. Commands to verify the success of each previous step | 2. Commands to verify the success of each previous step | ||
</pre> | </pre> | ||
== Role Playing == | == Role Playing == | ||
| Line 176: | Line 181: | ||
× List important sentences in bullet points | × List important sentences in bullet points | ||
* Meeting follow-up items | * Meeting follow-up items | ||
=== Fact-check === | |||
When fact-checking online rumors, it is recommended to include a source citation instruction in the prompt to improve the verifiability of the content. | |||
A common practice is to append the following instruction to the prompt: | |||
<pre> | |||
Please use footnote-style citations to mark all referenced online sources, and list all cited sources at the end of the article. | |||
Once this format is applied, each online reference within the body text will be marked with a numbered citation (e.g. `[1]`, `[2]`), with all corresponding sources compiled at the end of the article for readers to verify further. | |||
**See Also** | |||
- Fact-checking | |||
- Citation formats | |||
- Footnotes and endnotes | |||
**References** | |||
- *(List relevant sources here)* | |||
</pre> | |||
=== Plain English Writing === | === Plain English Writing === | ||
| Line 267: | Line 291: | ||
[Detailed rules for generating clickbait headlines follow...] | [Detailed rules for generating clickbait headlines follow...] | ||
=== | === News Quality Assessment === | ||
Prompt | |||
Prompt | |||
<pre> | <pre> | ||
Please help me analyze whether a news report meets rigorous journalistic standards. | |||
The topic this time is a breaking news report on <TITLE>, and I want to clarify the following points: | |||
(1) Primary question: Does this news report cite multiple sources? Does it have sufficient fact-checking and balance? | |||
(2) Extended question: Besides the number of sources, what other news quality indicators (such as background context, objectivity, structural completeness) have not been adequately considered? Is it possible that the reporting format (such as breaking news) might affect the judgment of its qualification? | |||
Please help me: | |||
- Establish a news quality assessment framework (can be presented in table format) | |||
- Apply this framework to analyze this report | |||
</pre> | |||
== English Learning == | == English Learning == | ||
=== English Conversation Practice === | === English Conversation Practice === | ||
| Line 447: | Line 456: | ||
</pre> | </pre> | ||
== De-identification of photo == | |||
Prompt | |||
<pre> | |||
Please transform this uploaded image into a vibrant anime watercolor style. Use soft brushstrokes and a warm color palette. Important: Do not retain the specific facial features of the people in the photo; instead, reimagine them as generic anime characters with clear eyes and expressions that match the scene. | |||
</pre> | |||
Model: Gemini 2.5 Flash Image (Nano Banana) | |||
== System Prompt == | == System Prompt == | ||