A 364M Hebrew decision encoder on NeoDictaBERT: give it a state, a question and 2 to 7 free-text options and it returns a probability for each option in one forward pass, about 70 ms on a laptop CPU. Hebrew only.
An independent Jev-shaped reproduction on DeBERTa-v3-large. One state and any number of choice, score and yes/no questions go in, and a calibrated distribution per question comes out of one forward pass. Public gold labels only.
Decides
choice, classify, route
choice, score, noul, classify
Architecture
nativ
open-jev
Fine-tuned from
dicta-il/neodictabert
microsoft/deberta-v3-large
License
cc-by-4.0
apache-2.0
Availability
Open weights
Open weights
Hosted by
—
—
Input price
—
—
Decision accuracy
96.2%
85.4%
Calibration error
0.010
0.022
Valid action rate
—
—
Median latency
—
28 ms
p95 latency
—
—
Evaluation suite
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 nativ-he-decision and open-jev-deberta-v3-large?
nativ-he-decision is from Yoav Pinto and open-jev-deberta-v3-large from Kotoba Labs. Both have open weights you can download and run. Both answer choice and classify questions. Only nativ-he-decision answers route. Only open-jev-deberta-v3-large answers score and noul. nativ-he-decision reads up to 1K tokens of state, against 512 tokens for open-jev-deberta-v3-large. nativ-he-decision is the smaller model, at 364M parameters to 434M. nativ-he-decision is licensed cc-by-4.0; open-jev-deberta-v3-large, apache-2.0.
Which is more accurate, nativ-he-decision or open-jev-deberta-v3-large?
They report on different suites — nativ-he-decision 96.2% on nativ-bench user-intent task, 4 options (MASSIVE Hebrew, 2,973 items; the maker's own suite), open-jev-deberta-v3-large 85.4% on Kotoba held-out test, seen question types (banking77, SST-5, BoolQ) — so the numbers do not rank them. Test both on your own labelled examples.
Which is cheaper, nativ-he-decision or open-jev-deberta-v3-large?
nativ-he-decision: Free (open weights). open-jev-deberta-v3-large: Free (open weights). Open weights cost nothing per call beyond your own hardware.
Can I run nativ-he-decision or open-jev-deberta-v3-large locally?
Yes, both: systemone pull yoav-pinto/nativ-he-decision and systemone pull kotoba-labs/open-jev-deberta-v3-large download the weights.
nativ-bench user-intent task, 4 options (MASSIVE Hebrew, 2,973 items; the maker's own suite)
Kotoba held-out test, seen question types (banking77, SST-5, BoolQ)
