Skip to content
System One

Zero-shot classifier

A zero-shot classifier assigns text to labels it was never trained on, using the label descriptions you supply at call time rather than labelled examples. Jev's Choice question works this way: you define the categories in the request, and the model returns a probability distribution over them.

The term predates Jev by years. Encoder classifiers such as BERT and DeBERTa, and span models such as GLiNER, have been doing classification against labels supplied at inference time for a while, which is why the comparison came up straight after the launch on 2026-09-15.

agentpedia.codes reports that a Hacker News commenter called Jev “basically a zero-shot classifier” and that Diogo Almeida replied “exactly right!”. That exchange could not be confirmed in the Hacker News thread itself, so treat it as agentpedia.codes’ account rather than a direct citation from HN.

Where the label stops fitting is the request shape. A zero-shot classifier usually means one labelling task per call. A Jev request carries a map of independent questions evaluated in parallel, mixing Choice with Score and Noul, and returns a calibrated probability distribution for each. One support ticket can be routed to a department, scored 1.30 for bug severity and flagged at 0.99 on “Is the customer asking for a human agent?” in a single call.

Whether that adds up to a new model class is the actual argument. Is Jev just a classifier sets out both sides, and choice covers the primitive itself.

Related terms