Intent Miss-classification with number intent

[engahmed1190]

Describe the bug
I am having a problem with the new NLU pipeline,

Numbers are misclassified when tested alone. previously it was working fine.

Old NLU Pipeline

pipeline:
  - name: WhitespaceTokenizer
  - name: CountVectorsFeaturizer
  - name: EmbeddingIntentClassifier
  - BILOU_flag: true
    name: CRFEntityExtractor
    features:
      - [low, title, upper]
      - [low, bias, prefix5, prefix2, suffix5, suffix3, suffix2, upper, title, digit, pattern]
      - [low, title, upper]
  - name: rasa_addons.nlu.components.gazette.Gazette
  - name: EntitySynonymMapper

Please help me to address this issue. Thanks in advance

[pheel]

What is the expected/previous behavior? If digits only appear under that one intent it's very likely the trained model will do this.

[engahmed1190]

Previously it was classified as none which is correct but Currently it is mapped to the intent that contains number. How may I address this, any workarounds