Automated README summary
Automatically summarized from the README on 4 Nov 2026; not a run by jevbooks. Run instructions and author-reported outcomes still need verification.
What it is
Valen-Decision is an open, lightweight multimodal decision model for text, images and video. It evaluates inputs against task instructions and returns probabilities over supplied candidates, giving software a structured decision interface. Includes training and inference code.
How it uses Jev
Inspired by Jev, Valen evaluates text, images and video against task instructions and returns probabilities over supplied candidates. Questions are typed (predicate, choice, score); answers carry probabilities. The README shows API usage returning a probability, but does not detail internal Jev decision logic.
Primitives:choicescore
Technique worth stealing
Calibrated routing: use probabilities to decide which tickets to automate versus send to human review, adding a 'none of these' option.
Try it
Call the free API with requests, or use the web chat; no download or GPU required.
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-1.14.0- Project signals
jev-1.14.0- Category
jev-1.14.0- Pattern tags
jev-1.14.0
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 projectnot found in the excerpt0.07
- Shows a System One patternnot found in the excerpt0.08
- Describes handling uncertaintynot found in the excerpt0.03
- Reports measured numbersnot found in the excerpt0.04
- Is a runnable toolFound in material0.93
- Top-10 learning shortlist signalnot found in the excerpt0.04
- Is a replica of the modelFound in material0.63
- Solves a common problem (0–2)score on a 0–2 scale1.68 / 2
- About JevFound in material0.78
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 4 Nov 2026.