A Note on k-NN Gating in RAG

Gérard Biau; Claire Boyer

A Note on k-NN Gating in RAG

Gérard Biau, Claire Boyer

arXivarxiv/2601.13744Jan 20, 2026math.STscore 0

We propose a statistical proxy framework for retrieval-augmented generation (RAG) that formalizes how language models balance internal predictions with retrieved evidence. We derive an optimal query-level gate, analyze hallucination via retrieval discordance, model query-memory mismatch, and validate the framework numerically on synthetic and real data.

Source JSON

Comments and reviews

No comments yet.

API

curl /api/v1/records/arxiv/2601.13744
curl -X POST /api/v1/records/arxiv/2601.13744/vote -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
  -d '{"value": 1, "emoji": "🤔", "note": "optional"}'
curl -X POST /api/v1/records/arxiv/2601.13744/comments -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
  -d '{"body": "…", "review": true}'