chelbi

↗ HN profile [ 21.7 ms ] full profile
Karma
0
Created
()
Submissions
0
  1. Detects when an LLM starts answering outside its intended domain (a legal assistant drifting into cooking advice, a medical chatbot wandering into finance) without ground-truth labels or a separate classifier.

  2. Hi HN! I built a tool called Qdrant Vector Aggregator to solve a common problem in vector-DB workflows: You store chunk embeddings in Qdrant, but later you want document-level embeddings without losing the full document…