Rasa Pro - connection with OpenAI in a chat where the user is asked for their name, address, email, and phone number, is generating over 3,000 tokens each time. For such a short conversation, it should be a maximum of 100 tokens. Could this be related to the chat history being automatically sent to OpenAI? Each chat is a separate session. Why are so many tokens being consumed? Currently, one short CALM conversation costs 10 cents if set type: rephrase = false and $0,25 if set type: rephrase = true. If you wanted to use Rasa for a voice bot, the cost of a simple registration request would be a few dollars. (OpenAi)