Automated README summary
Automatically summarized from the README on 16 Oct 2026; not a run by jevbooks. Run instructions and author-reported outcomes still need verification.
What it is
JevAlt provides three 4B open-weights language models for English, Turkish, and German that choose between options and return a probability for each, for routing, policy checks, and typed decisions on your own hardware.
How it uses Jev
JevAlt implements the Jev API (POST /v1/systemone). Given a state and typed questions (choice, score, noul), the models return calibrated probabilities for each option. The server allows existing Jev clients to use these models for decisions.
Primitives:choicescorenoul
Technique worth stealing
Fine-tune Intern-Decision-4B with LoRA to output per-option probabilities for typed decisions.
Try it
pip install "jevalt[serve,gguf] @ git+https://github.com/mertkayacs/jevalt" && jevalt serve
Model judgments from README excerpts
Project signals use repository metadata, an existing summary and the first 700 README characters, not the full project. “Not found” means not detected in that material; it does not establish absence. These are the original judgments with recorded models and dates, not source audits or run results.
Models & evaluation dates
- Inclusion gate
jev-1.13.0- Project signals
jev-1.13.0- Category
jev-1.13.0- Pattern tags
jev-1.13.0
Evidence
These signals use repository metadata, an existing summary and at most 700 cleaned README characters. ≥ 0.60 means reported in that excerpt; ≤ 0.40 means not found there. The middle band is inconclusive. None is a source-code audit or a test result. The runnable-tool question excludes research experiments; the recommendation question asks for a learning shortlist, not a quality verdict.
- Jev is core to the projectFound in material0.66
- Shows a System One patternnot found in the excerpt0.11
- Describes handling uncertaintynot found in the excerpt0.04
- Reports measured numbersnot found in the excerpt0.05
- Is a runnable toolFound in material0.84
- Top-10 learning shortlist signalunclear0.40
- Is a replica of the modelFound in material0.91
- Solves a common problem (0–2)score on a 0–2 scale1.88 / 2
- About JevFound in material0.95
no-judgment band (0.40–0.60)
no-judgment band (0.40–0.60)
The Jev model and date for each judgment stage are listed above. Card written by DeepSeek V4.1 Flash from the README on 16 Oct 2026.