Automated README summary
Automatically summarized from the README on 25 Sept 2026; not a run by jevbooks. Run instructions and author-reported outcomes still need verification.
What it is
JevK5 is an Apache-2.0 open-weight model and Python runtime for typed decisions. Given a state (ticket, log, policy, diff) and a yes/no, choice, or score question, it returns a probability per option for developers who want Jev-style decisions on their own GPU.
How it uses Jev
JevK5 is itself the model, not a user of Jev. It answers `noul` (yes/no), `choice`, and `score` questions by a softmax over answer-letter logits in one forward pass, with no generated tokens. The server accepts the TypeSafe-style `/v1/systemone` request shape; each question is evaluated separately and requests are serialized on one GPU.
Primitives:choicescorenoul
Technique worth stealing
SemIf's option-logit readout: softmax over answer letters' next-token logits with one temperature, plus one CUDA graph per padded input length.
Try it
pip install "jevk5[fast] @ git+https://github.com/allebee/jevk5@v0.3.0", then JevK5("alibiserikbay/JevK5").decide(...)
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 material0.91
- Shows a System One patternnot found in the excerpt0.15
- Describes handling uncertaintynot found in the excerpt0.04
- Reports measured numbersFound in material0.97
- Is a runnable toolFound in material0.87
- Top-10 learning shortlist signalnot found in the excerpt0.21
- Is a replica of the modelFound in material0.97
- Solves a common problem (0–2)score on a 0–2 scale1.14 / 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 25 Sept 2026.