RAG human handoff with Jev
Retrievers hope. After retrieval, Jev marks relevance, contradiction, or injection; code keeps, flags, or drops passages before a generator sees them.
This unofficial page is the human handoff slice of the RAG passage decisions pack. Intent: apply the Jev (TypeSafe System One) decision model to RAG passage decisions human handoff. Primary search language: RAG Jev human handoff. Confirm patterns on docs.typesafe.ai. This site does not sell, issue, or proxy TypeSafe keys. Use a credential you already have from the console or a documented gateway.
Independent angle (cover ≠ clone): Filter/rerank recipes with failure modes; cite vs generate boundary. We cover the intent, not a rival rerank-passages-score URL tree.
RAG use-case context
Handoff is a first-class outcome for RAG passage decisions, not a failure of Jev. When the retrieval filter cannot auto-act, a human sees a packed query + retrieved passage — not a chat transcript. We do not ask Jev to write the reviewer essay.
Hub: Classifying RAG passages. Compare, when the other tool is the real job: RAG pipelines.
Human Handoff inputs
Humans should see what the model saw (filtered), not the warehouse dump you correctly refused to POST:
{
"query": "What is the refund window for pro plans?",
"passage": { "id": "doc-88#p3", "text": "Pro subscribers may request a refund within 14 days." },
"corpus": { "trust": "internal_kb" }
}Decision signals and actions
| Signal | Typical reason enum (you name it) |
|---|---|
| Top passages all sit near 0.5 relevance | weak_retrieval |
| Contradiction Noul high among keepers | doc_conflict |
| Injection Noul mid-band | passage_review |
| Empty retriever | skip_jev — do not invent a passage |
These are application outcomes next to HTTP 200, not invented TypeSafe status codes.
Send the reviewer:
- query + passage.id + passage.text
- relevant / contradiction / injection values
- whether code dropped or kept
- what the generator was allowed to see
Do not treat a Noul of 0.5 as a “medium” RAG passage decisions score — it means yes and no are equally likely. Conjunctions stay in your code.
Guardrails and escalation
Do not page humans on a single Score unless your conjunction says so. Pair with confidence thresholds. TypeSafe’s confidence-gated examples use a lower bar for recoverable reads than for irreversible actions. Those numbers are illustrations. For RAG passage decisions, treat feed_to_answerer as the high bar (showing a passage to a customer-facing answerer). Tune on labels — see offline evaluation.
Evaluation and rollout notes
Write the gold label back into the offline set. That is how floors move. Pin jev-1.13.0 (the versioned id) after you fit thresholds. jev-latest and the marketing line jev-1.13 can move. Log the response model. TypeSafe’s published list price for jev-1.13 is $0.042 per million input tokens (vendor claim — confirm on the models page); output tokens are free on that same page. Unused distractors still bill as input.
Official Python and JavaScript SDKs read TYPESAFE_API_KEY and retry documented 429/529. This site does not sell, issue, or proxy TypeSafe keys. Use a credential you already have from the console or a documented gateway.
Pack map
| Slice | Page |
|---|---|
| Graph and primitives | decision workflow |
What may enter state |
input contracts |
| What to gather first | evidence collection |
| Atomic rules | policy checks |
| Act / review / abstain | confidence thresholds |
| Reviewer payload | you are here |
| What to persist | audit trail |
| How it breaks | failure modes |
| Labeled replay | evaluation |
| Shadow → canary | production rollout |
FAQ
Is handoff a Jev failure? No. It is a first-class outcome. Abstention is “no auto action”; handoff is the queue you send that case to (glossary).
Should Jev draft the reviewer note? No. Send structured answers. Generation is the wrong job.
Where is the rest of the RAG pack? Start with RAG confidence thresholds and RAG audit trail. Cluster hub: Use cases.
Should Jev generate the RAG answer? No. Classify or score passages; another model (or extractive code) writes. That is the cite-vs-generate boundary.
Do we publish rerank lifts? No. TypeSafe’s cookbooks may show measurements — treat those as vendor figures and re-run on your corpus.
What this page does not claim
- No invented top-1 / top-10 lifts.
- Not a vector database.
- Not official TypeSafe.
- Official TypeSafe status, or that jev.pro issues API keys.
- That a schema-constrained answer is automatically factually correct.
Disclaimer
This is an independent unofficial site and is not affiliated with TypeSafe AI; official documentation is available at https://docs.typesafe.ai.
Primary documentation: https://docs.typesafe.ai. Hub: Use cases.
Sources
Public TypeSafe or adjacent documentation only. No private claims.