Reference
Spam LLM properties
The AI model to use for the LLM classifier.
On this page
LLM Classifier
Model
The AI model to use for the LLM classifier.
One of your AI Models, chosen from a list. Name in the API: modelId.
Temperature
The temperature to use for the LLM classifier.
A number. Default: 0.5. Name in the API: temperature.
Prompt
The prompt to use for the LLM classifier.
Text, over several lines if needed. Name in the API: prompt.
Response Format
Separator
The separator character used to parse the LLM response.
Text. Default: ,. Name in the API: separator.
Category Index
The position of the category field in the LLM response.
A number. Default: 0. Name in the API: responsePosCategory.
Confidence Index
The position of the confidence field in the LLM response.
A number. Default: 1. Name in the API: responsePosConfidence.
Explanation Index
The position of the explanation field in the LLM response.
A number. Default: 2. Name in the API: responsePosExplanation.
Categories
The expected categories in the LLM response.
A list. Each one: text. Default: Commercial, Harmful, Legitimate, Unsolicited. Name in the API: categories.
Confidence
The expected confidence levels in the LLM response.
A list. Each one: text. Default: High, Low, Medium. Name in the API: confidence.