Automated README summary
Automatically summarized from the README on 3 Nov 2026; not a run by jevbooks. Run instructions and author-reported outcomes still need verification.
What it is
Understudy is a drop-in proxy for System One models like Jev, Rune, and Clef. It forwards requests to the lead model, trains small CPU models on the lead's answers, and answers repeated narrow questions itself once they graduate.
How it uses Jev
Understudy proxies Jev's typed questions (noul, choice, score) over any JSON state. It forwards each request to Jev, stores the answer as a training example, and trains understudents to predict Jev's answer. Once a question graduates, Understudy answers it locally with calibrated confidence, auditing against Jev and demoting on drift.
Primitives:choicescorenoul
Technique worth stealing
Distill a fixed, repeated question from lead answers using grouped out-of-fold CV and Wilson-bound graduation, then audit and demote on drift.
Try it
Run `understudy serve --lead-url <url>` and set `TYPESAFE_BASE_URL=http://127.0.0.1:8787`.
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 excerpt0.23
- Shows a System One patternnot found in the excerpt0.18
- Describes handling uncertaintynot found in the excerpt0.03
- Reports measured numbersnot found in the excerpt0.03
- Is a runnable toolFound in material0.78
- Top-10 learning shortlist signalnot found in the excerpt0.21
- Is a replica of the modelFound in material0.85
- Solves a common problem (0–2)score on a 0–2 scale1.02 / 2
- About JevFound in material0.98
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 3 Nov 2026.