The AI task is a powerful feature that allows you to perform any AI task in the runtime.
Types of Response Sources:
In Studio, Response Source include two options one from AI
and second from Knowledgebase
.
1. AI:
AI uses the pretrained (their built-in) knowledge or any additional context you provide to generate responses.
- Previous Chat Context: Specify how many previous messages the AI should consider for context.
- Prompt: Add instructions for the AI to define how it should respond.
2. Knowledgebase:
The Knowledgebase AI uses your training data to answer to the users question.
- Previous Chat Context: Specify how many previous messages the AI should consider for context.
- Question: Enter the userβs query to retrieve specific information.
- Prompt: Add extra context or instructions for how the AI should handle the query.
- Restrictions: Define rules or limits on how the AI generates responses.
Note: For most tasks, up to 5 messages is sufficient. Adding too many messages may reduce response quality.
Common Settings for AI and Knowledgebase
These settings are common for both AI and Knowledgebase data sources within YourGPT Chatbot Studio:
-
Send to Chat: The generated response is delivered to the chat interface for user interaction and engagement.
-
Store Into: The Basic information is stored across various visitor, session, Contact and Flow parameters, including visitor details, session information, and bot-related data.
Model Configuration Settings
The configuration settings for both AI and Knowledgebase data sources are standardized, offering flexibility and customization:
- OpenAI
- Anthropic
- Mistral
- Gemini
Choose the AI model for response generation:
- GPT-3.5-turbo
- GPT-3.5-turbo-1106,
- GPT-4
- GPT-4o
- Claude-3.5-Sonnet
- Claude-3-Opus
- Gemini-1.5-Flash
- Mistral-2-Large
Temperature is a parameter that controls the creativity & randomness of the AI modelβs output.
0-1 (adjustable)
Max Tokens is the maximum number of tokens that the AI model can generate. Sets the maximum response length (default: 1024).
1-2048 (adjustable)
Learn more about Tokens, Max Tokens, Context Limits, Knowledge Nodes, and Temperature.