Hello, I am using fallback policies as you can see here.
I set the value for both nlu an core to 0.5 yet I get this when I ran it.
Your NLU fallback is 0.5, and it classified goodbye with confidence 0.36 – it should fall back in this case – I don’t see the problem. Then it wants to run utter_fallback because it is not confident enough in action prediction (core threshold). When it says the bot wants to run utter_fallback, you should tell it instead to do the correct action of what the bot should do after the intent that you corrected.
1 Like