I had been trying out Rasa X for interactive learning as it is a nightmare on the CLI. I have built Rasa X locally for this reason and I am trying out a few possibilities. However, When I go for the talk to my bot option everything is normal until I correct an action. There are two things happening.
The session restarts. However, the story is not discarded which is a good thing.
However, there are a few injections in terms of steps in the story. This is not what I would ideally want.
Hey. I won’t be able to give you more than the screenshots here. However, what I have experienced is that basically, rasa x goes back to a random step above (and) or a few intermediate steps disappear.
My impression is that in the story the correction_action is triggered and that after this happened something unexpected happened. However, when I look at the 2nd screenshot I don’t see the correction_action listed.
Just for my understanding; what is this action supposed to do? Is it a custom action you’ve implemented? Is there a reason why the action isn’t called action_correct?
Ah wait. I think I may have misinterpreted. When you refer to correction_action you’re not talking a custom action but something you’ve done from the UI? Are you able to share a GIF perhaps?
