Skip to content

Instantly share code, notes, and snippets.

@martinbowling
Last active May 5, 2024 22:57
Show Gist options
  • Save martinbowling/316fef6d383e00c7fb529d0fdcdff8ed to your computer and use it in GitHub Desktop.
Save martinbowling/316fef6d383e00c7fb529d0fdcdff8ed to your computer and use it in GitHub Desktop.
One Shot Web Spam Classifier Claude 3 Opus
<prompt>
Please analyze the following 10 examples of web spam content and 10 examples of good web content:
Web Spam Examples:
<spam_example1>
[URL, title, meta description, h1, h2, h3, h4 for web spam example 1]
</spam_example1>
<spam_example2>
[URL, title, meta description, h1, h2, h3, h4 for web spam example 2]
</spam_example2>
...
<spam_example10>
[URL, title, meta description, h1, h2, h3, h4 for web spam example 10]
</spam_example10>
Good Web Content Examples:
<good_example1>
[URL, title, meta description, h1, h2, h3, h4 for good web content example 1]
</good_example1>
<good_example2>
[URL, title, meta description, h1, h2, h3, h4 for good web content example 2]
</good_example2>
...
<good_example10>
[URL, title, meta description, h1, h2, h3, h4 for good web content example 10]
</good_example10>
For the web spam examples, analyze each component and identify common patterns:
URL: Look for unusual domain names, excessive keywords, or suspicious TLDs
Title: Check for keyword stuffing, excessive length, or misleading titles
Meta Description: Identify keyword stuffing, lack of relevance, or excessive length
Headings (H1, H2, H3, H4): Look for keyword stuffing, lack of hierarchy, or irrelevant headings
For the good web content examples, analyze each component and identify best practices:
URL: Look for concise, relevant, and readable URLs
Title: Check for concise, relevant, and engaging titles that accurately describe the page content
Meta Description: Identify well-written, relevant, and compelling descriptions that encourage clicks
Headings (H1, H2, H3, H4): Look for clear, relevant, and properly structured headings that enhance readability and SEO
After identifying the common patterns and best practices, analyze the target site:
<target_site>
[URL, title, meta description, h1, h2, h3, h4 for the target website to be analyzed]
</target_site>
Compare each component of the target site to the patterns found in the spam and good content examples:
URL: Determine if the URL is more similar to the spam or good content examples
Title: Assess whether the title exhibits characteristics of spam or good content
Meta Description: Evaluate if the meta description is more aligned with spam or good content practices
Headings (H1, H2, H3, H4): Analyze the headings and determine if they resemble spam or good content patterns
Provide the results in a table with this format:
| URL | Title | Spam Likelihood (1-10) | Reasoning |
In the "Spam Likelihood" column, rate the likelihood of the target site being spam on a scale from 1 to 10 (1 = very unlikely to be spam, 10 = almost certainly spam).
In the "Reasoning" column, provide a detailed explanation of how the target site's components compare to the spam and good content examples. Be specific and reference the key patterns and best practices you identified earlier. Explain how these comparisons contribute to the spam likelihood score you assigned.
</prompt>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment