Skip to content

Conversation

@zhuohangu
Copy link
Collaborator

concept_generator.py implements LLM-based concept generation with two modes:

  • DIRECT: generates a deduped list of “final concepts” from a list of queries.

  • TWO_STAGE [DEFAULT]: per-query concept generation → SentenceTransformer embeddings + KMeans clustering → LLM centroid generation.

It persists results to a JSON artifact (including intermediate outputs for TWO_STAGE) and includes a minimal QUEST runnable example.

@zhuohangu zhuohangu requested review from Copilot and mdr223 and removed request for Copilot December 17, 2025 20:38
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants