Phishing example fix

[efajardo-nv]

Description

• Update labels file to reflect actual model outputs: not_phishing and is_phishing
• Add AddScores stage to Python and CLI examples
• Remove FilterDetectionsStage because model output contains probabilities for both not_phishing and is_phishingone of which will always exceed threshold. This results in nothing ever being filtered out.

Closes #1208

By Submitting this PR I confirm:

• I am familiar with the Contributing Guidelines.
• When the PR is ready for review, new or existing tests cover these changes.
• When the PR is ready for review, the documentation is up to date with these changes.

[mdemoret-nv]

/merge