Benchmarks & research

JevTuner

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.

Repository ↗Pinned README ↗Pinned implementation ↗Pinned implementation ↗

Primitives: choice
Platform: Python
Added: Sep 22, 2026
Project created: Sep 21, 2026
GitHub stars: 13 · snapshot 2026-09-25

Source checked 2026-09-25 — opened the primary source directly.