Activechat Manual
  • What is Activechat?
  • New? Start here
    • The basics
    • Set up your first project
    • Install the chat widget
    • Upload the knowledge
    • Explore the CRM
    • AI-assisted live chat
      • How to set context for AI hints
    • Live chat mobile app
    • Build your first automation
  • Conversational AI
    • For Customer Service Teams
    • For Product Managers
    • For Innovation Teams
    • For Marketers
    • For e-commerce
    • For developers
  • Help Guides
    • Setting up your team
    • Managing conversations
      • Customer attributes
      • User tags and segments
      • Searching for specific users
      • Agent tags (live chat groups and queues)
      • Triggering live chat sessions from the bot
      • Notifications with the TRIGGER block
    • Managing knowledge
      • Uploading business data
      • Question answering and live chat hints
      • Fine-tuning the large language model
    • Natural language automation
    • Building automations visually
      • Customizing your welcome message
      • Adding new skills
      • Navigating skills
      • Copying skills and blocks
      • Handling errors
    • Improving your virtual agent
    • Using live chat AI hints
    • Customizing automatic website page messages
    • Tracking website actions
    • Facebook Ads automation
      • How to set up a Facebook ads bot
      • How to use buttons and quick replies in a Facebook ads chatbot
    • Lead generation
    • Zapier integrations
    • Customizing your project
      • How to customize the chat widget
      • How to customize the Facebook chat widget
      • How to change bot settings
    • Pricing guide
  • Fundamentals
    • Terminology
      • Intents and entities
      • Contexts
      • Skills and events
        • Built-in system skills
          • /start
          • /default
          • /_default_fallback
          • /_start_live_chat
          • /_page_visit
          • /_error
      • Conversation elements
        • Messages
        • Buttons
        • Quick replies
        • Galleries / carousels
    • Messaging channels
      • Website chat widget
        • Installation
        • Customization
        • Voice input
      • Chat widget landing page
      • Facebook Messenger
        • Connect your page
        • 24 hour rule
        • Message tags
        • Persistent menu
      • Telegram
      • Email
      • Twilio SMS automation
    • Intents and bot skills
    • Conversation insights
    • Grow tools
      • Landing pages
      • Messenger links and QR codes
    • Broadcasting
  • Visual builder reference
    • Sending messages
      • TEXT
      • LISTEN
      • IMAGE
      • MEDIA
      • GALLERY
      • FILE
      • EMAIL
      • SMS
      • LEAD
    • Triggering events
      • SEND
      • CATCH
      • TRIGGER
      • LIVE CHAT
    • Manipulating data
      • DATA
      • ADD TAG
      • REMOVE TAG
      • JSON
      • STATUS
      • VALIDATION
    • Conditional logic
      • SWITCH
    • Timers and delays
      • TIMER
      • WAITFOR
      • WAITUNTIL
    • E-commerce blocks
      • CATEGORY
      • PRODUCT
      • VARIATIONS
      • SIMILAR
      • UPSELLS
      • CROSSSELS
      • Shopping carts
        • ADD TO CART
        • UPDATE CART
        • SHOW CART
        • CLEAR CART
        • CREATE ORDER
    • Natural Language
      • NLP
    • System attributes
    • System events
  • Integrations
    • Google services
      • Connect your Google account
      • Google Sheets
        • Searching and updating Google Sheets data
        • Building galleries with Google Sheets data
      • Google Calendar
        • Searching for events
        • Creating and updating events
    • Shopify
    • WooCommerce
    • Dialogflow
      • Building an agent
      • Using entities
      • Slot filling
      • Context management
      • E-commerce NLP
Powered by GitBook
On this page
  • Fine-tuning through the conversation history
  • Fine-tuning from the live chat interface

Was this helpful?

  1. Help Guides
  2. Managing knowledge

Fine-tuning the large language model

PreviousQuestion answering and live chat hintsNextNatural language automation

Last updated 2 years ago

Was this helpful?

When you upload your business knowledge and perform the initial model training, it will correctly recognize and respond to around 70% of questions that customers may ask about your data. To improve this, you should spend some time fine-tuning the model. After fine-tuning, 90% or even 95% response quality is not unusual.

Fine-tuning can be done in the part of the interface or from the live chat hints panel.

To use fine-tuning, your project should be on the "Company" or "Enterprise" plan with AI engine enabled in the project settings.

Fine-tuning through the conversation history

When you open the conversation log for any user in your project, you will see small icons next to every message sent by the bot or human agent. These icons can be used to fine-tune the ML model, improving the quality of automated responses.

Click the green checkbox if you like the response, or red cross if you find it irrelevant or poorly formulated. When you mark response as "bad", a small pop-up will appear, asking you to enter the link to the source of the correct reponse. You can skip this, simply marking the response as irrelevant, or provide the model with more info to use in further conversations.

Fine-tuning from the live chat interface

Fine-tuning icons can also be found in the "AI hints" section of a live chat interface.

"Conversations"
Fine-tuning the AI model through conversations
Fine-tuning the AI model through live chat hints