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..

Interjection substitution incorrect?
 
 

:prepare “No, he is wrong” (with a comma) gives me what I’d expect:

1st sentence:

Original User Input: no, he is wrong
Tokenized into: no ,  he is wrong
Substituted (interjections ) into: ~no he is wrong
Actual used input: ~no(no ,)

2nd sentence:

Original User Input: he is wrong
Tokenized into: he is wrong
Actual used input: he is wrong


However, :prepare “No he is wrong” (without the comma) gives me this:

1st sentence:

Original User Input: no he is wrong
Tokenized into: no he is wrong
Substituted (interjections ) into: ~no he is wrong
Actual used input: ~no(no he)

2nd sentence:

Original User Input: is wrong
Tokenized into: is wrong
Actual used input: is wrong


..as you can see, the ‘he’ has gone missing from the start of the second sentence.  I can see that interjections.txt there is a line like this:

<no_he ~no+he

Does this mean that it should be replacing ‘no he’ with the interjection ‘~no’ and then the word ‘he’?  If so, where did the ‘he’ go?

 

 
  [ # 1 ]

a bug in substitution processing. Hard to erase. But I have just removed the substitions from LIVEDATA that map things like no_I to ~no+I

 

 
  login or register to react