“Do You Mean…?”: Fixing User Intent Without Annoying Them

by Mandy Jiang3 months ago
33

Autocomplete and intent detection systems can be helpful or infuriating. How often do models incorrectly “correct” users in a way that changes the intent? Can we evaluate intent preservation: when the model rewrites a query or instruction, does it keep the user’s goal? And can we design a method that improves clarification behavior without spamming questions?

Implemented:https://github.com/Hypogenic-AI/fix-user-intent-nlp-claudeImplemented:https://github.com/Hypogenic-AI/fix-user-intent-nlp-codexImplemented:https://github.com/Hypogenic-AI/fixing-user-intent-gemini
Chat

If you are inspired by this idea, you can reach out to the authors for collaboration or cite it:

@misc{jiang-do-you-mean-2026,
  author = {Jiang, Mandy},
  title = {“Do You Mean…?”: Fixing User Intent Without Annoying Them},
  year = {2026},
  url = {https://hypogenic.ai/ideahub/idea/86kBIv9R9Ypipm66AW2b}
}

Comments (0)

Please sign in to comment on this idea.

No comments yet. Be the first to share your thoughts!