Inform Intent with different entities that have the same values

Rasa Open Source
1

Hi all,

I am trying to figure out how to best deal with different entities that have the same values within an inform intent. For example, I have this nlu:

- intent: inform
  examples: |
    - I eat chocolate [often](chocolate) 
    - I [never](chocolate) eat chocolate
    - I [sometimes](chocolate) eat chocolate
    - [oftenn]{"entity": "chocolate", "value": "often"}
    ...
    -  [nver]{"entity": "chocolate", "value": "never"}
    ...
   - [ometimes]{"entity": "chocolate", "value": "sometimes"}
    - I [often](training) train
    - I [never](training) train
    - I [sometimes](training) train
    - [oftenn]{"entity": "training", "value": "often"}
    ...
    -  [nver]{"entity": "training", "value": "never"}
    ...
    - [ometimes]{"entity": "training", "value": "sometimes"}

The dots in between mean that I have many more examples for potential typos by the user. The problem is that if the bot classifies the intent inform it always uses chocolate as entity and not training. Any idea how I could circumvent this?