NewPaper Radar
Screen new research papers against personal interests with typed Jev judgments and publish a ranked reading feed.
Added to Jevfast
NewHow it uses Jev
Paper Radar sends each paper’s title, abstract and categories to Jev with separate Noul questions for interests, exclusions and release signals. Code combines the probabilities into reading bands and publishes pages and RSS. An optional text model summarizes only the selected papers.
What Jev decides
Source-defined request (illustrative)Example answers · not a recorded Jev response · Source ↗
Question 1 · s code
YOUR APP
INSTRUCTION
Does the abstract state that code, data or model weights are released or will be released?
STATE
A paper’s title, abstract and categories. This representative signal asks about public release claims in the abstract, not independently verified artifact availability.
JEV · NOUL
YesNo
App workflow
- Collect and deduplicate
The scheduled or manual run reads configured paper feeds and skips papers already handled.
- Judge each paper
A single Jev request carries the configured interest, exclusion and signal questions for that paper.
- Publish the reading feed
Code computes reading bands, stores decisions and builds pages and RSS; optional notifications and summaries follow.
Why it is interesting
The pipeline judges collected papers against explicit interests and keeps a decision audit, rather than asking a text generator to invent a digest. Code owns ranking, exclusions and calibration thresholds.