Automated README summary
Automatically summarized from the README on 21 Oct 2026; not a run by jevbooks. Run instructions and author-reported outcomes still need verification.
What it is
A TypeScript benchmark comparing Cloudflare clef-flash, TypeSafe Jev 1.14, Claude Haiku 5.5 and Amazon Rekognition on UnsafeBench's 10,146 images, reporting accuracy, latency and cost so others can rerun and rescore every number.
How it uses Jev
Jev receives the same two typed questions as the other prompted models: which of UnsafeBench's 11 categories the image relates to, and whether it is safe or unsafe. It returns a probability per answer; the top answer is Jev's own verdict, and a threshold chosen on the train split is applied to the test split for F1 and recall metrics.
Primitives:choice
Technique worth stealing
Compare models by recall at a fixed false-positive rate, not just F1, since scores mean different things across providers.
Try it
npm ci, fill .env keys, then npm run bench -- --model jev; npm run leaderboard or rescore.
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 projectnot found in the excerpt0.09
- Shows a System One patternnot found in the excerpt0.16
- Describes handling uncertaintynot found in the excerpt0.06
- Reports measured numbersnot found in the excerpt0.13
- Is a runnable toolnot found in the excerpt0.09
- Top-10 learning shortlist signalnot found in the excerpt0.19
- Is a replica of the modelnot found in the excerpt0.05
- Solves a common problem (0–2)score on a 0–2 scale1.95 / 2
- About JevFound in material0.97
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 21 Oct 2026.