NewJev Paper Map
A creator demo uses one 24-way Jev Choice per AI paper to build a searchable topic map, then separately evaluates the labels with Opus.
Added to Jevfast
NewHow it uses Jev
The creator screenshot states that every paper receives one 24-way Choice from Jev, returning a topic label and probability. The demo displays a topic map and searchable paper list for 1,000 papers, then separately compares those labels with an Opus judge. The post and screenshot establish this classification role, but do not disclose the exact request fields, question wording, option IDs or response-handling code.
Why it is interesting
It separates fast fixed-topic classification from an independent model judge, and exposes topic counts and per-paper confidence in a searchable map.