A training-free layer that turns an open LLM into a decision model. It reads typed choice, yes/no and score answers from one prefill, removes option-order bias with no labels and, from a few hundred labels, calibrates or fits a closed-form head. Weights stay untouched.
Ruoxi Qiu's decision adapter for a frozen Qwen3.8-27B: a 6.3M-parameter LoRA on the last four text layers plus a numerical prior. One shared state feeds isolated question branches; each Choice, Noul or Score question returns probabilities with no generated tokens.
Decides
choice, score, noul, classify, route
choice, noul, score
Architecture
anyjev
canopy-jev
Fine-tuned from
qwen/qwen3-8b
qwen/qwen3.8-27b
License
apache-2.0
apache-2.0
Availability
Open weights
Open weights
Hosted by
—
—
Input price
—
—
Decision accuracy
77.1%
87.4%
Calibration error
0.034
—
Valid action rate
—
—
Median latency
—
—
p95 latency
—
—
Figures are from each model’s manifest; accuracy and latency are what the publishers report, on their own suites and hardware. Add a third model.
Questions
What is the difference between anyjev and canopy-jev?
anyjev is from Nokia and canopy-jev from Ruoxi Qiu. Both have open weights you can download and run. Both answer choice, score and noul questions. Only anyjev answers classify and route. anyjev is the smaller model, at 8.0B parameters to 27B.
Which is more accurate, anyjev or canopy-jev?
They report on different suites — anyjev 77.1% on LocalLLaMA/typed-decisions (2,000 held-out decisions), L2 with 300 labels per question, canopy-jev 87.4% on JevBench public set (231 items; 202 correct), the maker's own fresh zero-shot run with the upstream scorer — so the numbers do not rank them. Test both on your own labelled examples.
Which is cheaper, anyjev or canopy-jev?
anyjev: Free (open weights). canopy-jev: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Can I run anyjev or canopy-jev locally?
Yes, both: systemone pull nokia/anyjev and systemone pull camellia86/canopy-jev download the weights.
Evaluation suite
LocalLLaMA/typed-decisions (2,000 held-out decisions), L2 with 300 labels per question
JevBench public set (231 items; 202 correct), the maker's own fresh zero-shot run with the upstream scorer
