Non-autoregressive System 1 decision engine. Typed choice, score and yes/no decisions over any text in a single forward pass, in 100+ languages, with a router that picks the right checkpoint per request.
LightJev
Train lightweight language backbones for typed decisions and candidate probabilities. CE/Brier training, evaluation, and an offline end-to-end demo.
Project facts
- Relationship to Jev
- Research
- Evidence
- Documented
- Language
- Python
- License
- Apache-2.0
- Origin
- Original repository
- Repository status
- Not archived
- Created
- 2026-09-18
- GitHub stars
- 2
- Evidence checked
- 2026-09-24T08:40:10.508Z
- Metadata checked
- 2026-09-24T08:40:10.508Z
- Check status
- current
Stars measure the whole repository, including work unrelated to Jev.
Evidence and scope
Documented records the linked documentation or source. JevHunt has not independently run or benchmarked this project.
**[Download LightJev-0.6B-v0.1]** — trained Qwen3-0.6B scoring weights, tokenizer, frozen gold-only data, training histories, and both CE/Brier evaluation reports. The published CE checkpoint was selected at step 250 using development CE before held-out results were inspected.
Evidence commit: dede59a91ff52305921a5e437d4cd1a39bab3be5
Discovered through: github-search.
How we review →