Benchmarks & research

JevTuner

Bookmark Keep this in your collection.

Leave a noteWhat would you try with this?

Only you can see your notes.

Source screenshot of JevTuner — SOURCE SCREENSHOTFull screenshot ↗

What it does

Train a Qwen-based closed-choice model by applying Brier loss directly to candidate-token probabilities; it explores Jev's decision-first idea but is not Jev.

How you can use it

For an app that sorts messages, start with examples already placed in the right folders. A developer can use JevTuner to teach an AI from those examples. When a new message arrives, the AI estimates how likely each folder is to be the right one. Your app gets a score for every folder, not a written reply.