AI Zone Admin Forum Add your forum

NEWS: Chatbots.org survey on 3000 US and UK consumers shows it is time for chatbot integration in customer service!read more..

unwanted interjection handling
 
 

TOPIC: ~work ()
t: Oh Fabro, how are you feeling 5?
  #! fine
  a: (![not] * [well ok fine health better]) Don’t you think it’s time you got a job?
  b: (~yes) So go get one. ^gambit(jobwhere)
  b: (~no) Well, I see you out late at night with your friends. You’re feeling better, go get a job. ^gambit(angry)
  #! still sick
  a: (~illness) Well, if you can go out and drink with your friends, you can work. Get out of here and go find a job. ^gambit(angry)

In the above, which is the start of the interaction, (work is the initial topic), if you answer

ok

it acts as if it had no match (matches a later gambit)

if you answer

I am doing ok

it gives the correct response

Is there a way to tune interjection splittling?

 

 

 
  [ # 1 ]

You can, of course, view what you are actually getting via
:prepare ok

You can shut down interjection processing entirely by changing $token. Or you can copy the LIVEDATA folder elsewhere, point to it on command line, and edit your own copies of the file to remove entries you don’t want. (moving it elsewhere means you can download revised CS and not get impacted).

 

 
  login or register to react