Zero-Shot Classification
Laya
Safetensors
GGUF
Hebrew
hebrew
decision-model
calibrated
scam-detection
routing
Instructions to use BrainboxAI/nitzotz with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Laya
How to use BrainboxAI/nitzotz with Laya:
# No code snippets available yet for this library. # To use this model, check the repository files and the library's documentation. # Want to help? PRs adding snippets are welcome at: # https://github.com/huggingface/huggingface.js
- Notebooks
- Google Colab
- Kaggle
Stage draft model card (Nitzotz v3 seed 2 numbers)
Browse files- .gitattributes +7 -0
- LICENSE +201 -0
- NOTICE +29 -0
- README.md +478 -0
- assets/banner_en.png +0 -0
- assets/banner_he.png +0 -0
- assets/bars_en.png +3 -0
- assets/bars_he.png +3 -0
- assets/calibration_en.png +3 -0
- assets/calibration_he.png +3 -0
- assets/radar_en.png +3 -0
- assets/radar_he.png +3 -0
- assets/ramzor_en.svg +37 -0
- assets/ramzor_he.svg +37 -0
- assets/speed_en.png +3 -0
- assets/speed_he.png +0 -0
- assets/tiles_en.png +0 -0
- assets/tiles_he.png +0 -0
.gitattributes
CHANGED
|
@@ -33,3 +33,10 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 33 |
*.zip filter=lfs diff=lfs merge=lfs -text
|
| 34 |
*.zst filter=lfs diff=lfs merge=lfs -text
|
| 35 |
*tfevents* filter=lfs diff=lfs merge=lfs -text
|
| 36 |
+
assets/bars_en.png filter=lfs diff=lfs merge=lfs -text
|
| 37 |
+
assets/bars_he.png filter=lfs diff=lfs merge=lfs -text
|
| 38 |
+
assets/calibration_en.png filter=lfs diff=lfs merge=lfs -text
|
| 39 |
+
assets/calibration_he.png filter=lfs diff=lfs merge=lfs -text
|
| 40 |
+
assets/radar_en.png filter=lfs diff=lfs merge=lfs -text
|
| 41 |
+
assets/radar_he.png filter=lfs diff=lfs merge=lfs -text
|
| 42 |
+
assets/speed_en.png filter=lfs diff=lfs merge=lfs -text
|
LICENSE
ADDED
|
@@ -0,0 +1,201 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
Apache License
|
| 2 |
+
Version 2.0, January 2004
|
| 3 |
+
http://www.apache.org/licenses/
|
| 4 |
+
|
| 5 |
+
TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
|
| 6 |
+
|
| 7 |
+
1. Definitions.
|
| 8 |
+
|
| 9 |
+
"License" shall mean the terms and conditions for use, reproduction,
|
| 10 |
+
and distribution as defined by Sections 1 through 9 of this document.
|
| 11 |
+
|
| 12 |
+
"Licensor" shall mean the copyright owner or entity authorized by
|
| 13 |
+
the copyright owner that is granting the License.
|
| 14 |
+
|
| 15 |
+
"Legal Entity" shall mean the union of the acting entity and all
|
| 16 |
+
other entities that control, are controlled by, or are under common
|
| 17 |
+
control with that entity. For the purposes of this definition,
|
| 18 |
+
"control" means (i) the power, direct or indirect, to cause the
|
| 19 |
+
direction or management of such entity, whether by contract or
|
| 20 |
+
otherwise, or (ii) ownership of fifty percent (50%) or more of the
|
| 21 |
+
outstanding shares, or (iii) beneficial ownership of such entity.
|
| 22 |
+
|
| 23 |
+
"You" (or "Your") shall mean an individual or Legal Entity
|
| 24 |
+
exercising permissions granted by this License.
|
| 25 |
+
|
| 26 |
+
"Source" form shall mean the preferred form for making modifications,
|
| 27 |
+
including but not limited to software source code, documentation
|
| 28 |
+
source, and configuration files.
|
| 29 |
+
|
| 30 |
+
"Object" form shall mean any form resulting from mechanical
|
| 31 |
+
transformation or translation of a Source form, including but
|
| 32 |
+
not limited to compiled object code, generated documentation,
|
| 33 |
+
and conversions to other media types.
|
| 34 |
+
|
| 35 |
+
"Work" shall mean the work of authorship, whether in Source or
|
| 36 |
+
Object form, made available under the License, as indicated by a
|
| 37 |
+
copyright notice that is included in or attached to the work
|
| 38 |
+
(an example is provided in the Appendix below).
|
| 39 |
+
|
| 40 |
+
"Derivative Works" shall mean any work, whether in Source or Object
|
| 41 |
+
form, that is based on (or derived from) the Work and for which the
|
| 42 |
+
editorial revisions, annotations, elaborations, or other modifications
|
| 43 |
+
represent, as a whole, an original work of authorship. For the
|
| 44 |
+
purposes of this License, Derivative Works shall not include works
|
| 45 |
+
that remain separable from, or merely link (or bind by name) to the
|
| 46 |
+
interfaces of, the Work and Derivative Works thereof.
|
| 47 |
+
|
| 48 |
+
"Contribution" shall mean any work of authorship, including the
|
| 49 |
+
original version of the Work and any modifications or additions
|
| 50 |
+
to that Work or Derivative Works thereof, that is intentionally
|
| 51 |
+
submitted to Licensor for inclusion in the Work by the copyright owner
|
| 52 |
+
or by an individual or Legal Entity authorized to submit on behalf of
|
| 53 |
+
the copyright owner. For the purposes of this definition, "submitted"
|
| 54 |
+
means any form of electronic, verbal, or written communication sent
|
| 55 |
+
to the Licensor or its representatives, including but not limited to
|
| 56 |
+
communication on electronic mailing lists, source code control systems,
|
| 57 |
+
and issue tracking systems that are managed by, or on behalf of, the
|
| 58 |
+
Licensor for the purpose of discussing and improving the Work, but
|
| 59 |
+
excluding communication that is conspicuously marked or otherwise
|
| 60 |
+
designated in writing by the copyright owner as "Not a Contribution."
|
| 61 |
+
|
| 62 |
+
"Contributor" shall mean Licensor and any individual or Legal Entity
|
| 63 |
+
on behalf of whom a Contribution has been received by Licensor and
|
| 64 |
+
subsequently incorporated within the Work.
|
| 65 |
+
|
| 66 |
+
2. Grant of Copyright License. Subject to the terms and conditions of
|
| 67 |
+
this License, each Contributor hereby grants to You a perpetual,
|
| 68 |
+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
| 69 |
+
copyright license to reproduce, prepare Derivative Works of,
|
| 70 |
+
publicly display, publicly perform, sublicense, and distribute the
|
| 71 |
+
Work and such Derivative Works in Source or Object form.
|
| 72 |
+
|
| 73 |
+
3. Grant of Patent License. Subject to the terms and conditions of
|
| 74 |
+
this License, each Contributor hereby grants to You a perpetual,
|
| 75 |
+
worldwide, non-exclusive, no-charge, royalty-free, irrevocable
|
| 76 |
+
(except as stated in this section) patent license to make, have made,
|
| 77 |
+
use, offer to sell, sell, import, and otherwise transfer the Work,
|
| 78 |
+
where such license applies only to those patent claims licensable
|
| 79 |
+
by such Contributor that are necessarily infringed by their
|
| 80 |
+
Contribution(s) alone or by combination of their Contribution(s)
|
| 81 |
+
with the Work to which such Contribution(s) was submitted. If You
|
| 82 |
+
institute patent litigation against any entity (including a
|
| 83 |
+
cross-claim or counterclaim in a lawsuit) alleging that the Work
|
| 84 |
+
or a Contribution incorporated within the Work constitutes direct
|
| 85 |
+
or contributory patent infringement, then any patent licenses
|
| 86 |
+
granted to You under this License for that Work shall terminate
|
| 87 |
+
as of the date such litigation is filed.
|
| 88 |
+
|
| 89 |
+
4. Redistribution. You may reproduce and distribute copies of the
|
| 90 |
+
Work or Derivative Works thereof in any medium, with or without
|
| 91 |
+
modifications, and in Source or Object form, provided that You
|
| 92 |
+
meet the following conditions:
|
| 93 |
+
|
| 94 |
+
(a) You must give any other recipients of the Work or
|
| 95 |
+
Derivative Works a copy of this License; and
|
| 96 |
+
|
| 97 |
+
(b) You must cause any modified files to carry prominent notices
|
| 98 |
+
stating that You changed the files; and
|
| 99 |
+
|
| 100 |
+
(c) You must retain, in the Source form of any Derivative Works
|
| 101 |
+
that You distribute, all copyright, patent, trademark, and
|
| 102 |
+
attribution notices from the Source form of the Work,
|
| 103 |
+
excluding those notices that do not pertain to any part of
|
| 104 |
+
the Derivative Works; and
|
| 105 |
+
|
| 106 |
+
(d) If the Work includes a "NOTICE" text file as part of its
|
| 107 |
+
distribution, then any Derivative Works that You distribute must
|
| 108 |
+
include a readable copy of the attribution notices contained
|
| 109 |
+
within such NOTICE file, excluding those notices that do not
|
| 110 |
+
pertain to any part of the Derivative Works, in at least one
|
| 111 |
+
of the following places: within a NOTICE text file distributed
|
| 112 |
+
as part of the Derivative Works; within the Source form or
|
| 113 |
+
documentation, if provided along with the Derivative Works; or,
|
| 114 |
+
within a display generated by the Derivative Works, if and
|
| 115 |
+
wherever such third-party notices normally appear. The contents
|
| 116 |
+
of the NOTICE file are for informational purposes only and
|
| 117 |
+
do not modify the License. You may add Your own attribution
|
| 118 |
+
notices within Derivative Works that You distribute, alongside
|
| 119 |
+
or as an addendum to the NOTICE text from the Work, provided
|
| 120 |
+
that such additional attribution notices cannot be construed
|
| 121 |
+
as modifying the License.
|
| 122 |
+
|
| 123 |
+
You may add Your own copyright statement to Your modifications and
|
| 124 |
+
may provide additional or different license terms and conditions
|
| 125 |
+
for use, reproduction, or distribution of Your modifications, or
|
| 126 |
+
for any such Derivative Works as a whole, provided Your use,
|
| 127 |
+
reproduction, and distribution of the Work otherwise complies with
|
| 128 |
+
the conditions stated in this License.
|
| 129 |
+
|
| 130 |
+
5. Submission of Contributions. Unless You explicitly state otherwise,
|
| 131 |
+
any Contribution intentionally submitted for inclusion in the Work
|
| 132 |
+
by You to the Licensor shall be under the terms and conditions of
|
| 133 |
+
this License, without any additional terms or conditions.
|
| 134 |
+
Notwithstanding the above, nothing herein shall supersede or modify
|
| 135 |
+
the terms of any separate license agreement you may have executed
|
| 136 |
+
with Licensor regarding such Contributions.
|
| 137 |
+
|
| 138 |
+
6. Trademarks. This License does not grant permission to use the trade
|
| 139 |
+
names, trademarks, service marks, or product names of the Licensor,
|
| 140 |
+
except as required for reasonable and customary use in describing
|
| 141 |
+
the origin of the Work and reproducing the content of the NOTICE file.
|
| 142 |
+
|
| 143 |
+
7. Disclaimer of Warranty. Unless required by applicable law or
|
| 144 |
+
agreed to in writing, Licensor provides the Work (and each
|
| 145 |
+
Contributor provides its Contributions) on an "AS IS" BASIS,
|
| 146 |
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
|
| 147 |
+
implied, including, without limitation, any warranties or conditions
|
| 148 |
+
of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
|
| 149 |
+
PARTICULAR PURPOSE. You are solely responsible for determining the
|
| 150 |
+
appropriateness of using or redistributing the Work and assume any
|
| 151 |
+
risks associated with Your exercise of permissions under this License.
|
| 152 |
+
|
| 153 |
+
8. Limitation of Liability. In no event and under no legal theory,
|
| 154 |
+
whether in tort (including negligence), contract, or otherwise,
|
| 155 |
+
unless required by applicable law (such as deliberate and grossly
|
| 156 |
+
negligent acts) or agreed to in writing, shall any Contributor be
|
| 157 |
+
liable to You for damages, including any direct, indirect, special,
|
| 158 |
+
incidental, or consequential damages of any character arising as a
|
| 159 |
+
result of this License or out of the use or inability to use the
|
| 160 |
+
Work (including but not limited to damages for loss of goodwill,
|
| 161 |
+
work stoppage, computer failure or malfunction, or any and all
|
| 162 |
+
other commercial damages or losses), even if such Contributor
|
| 163 |
+
has been advised of the possibility of such damages.
|
| 164 |
+
|
| 165 |
+
9. Accepting Warranty or Additional Liability. While redistributing
|
| 166 |
+
the Work or Derivative Works thereof, You may choose to offer,
|
| 167 |
+
and charge a fee for, acceptance of support, warranty, indemnity,
|
| 168 |
+
or other liability obligations and/or rights consistent with this
|
| 169 |
+
License. However, in accepting such obligations, You may act only
|
| 170 |
+
on Your own behalf and on Your sole responsibility, not on behalf
|
| 171 |
+
of any other Contributor, and only if You agree to indemnify,
|
| 172 |
+
defend, and hold each Contributor harmless for any liability
|
| 173 |
+
incurred by, or claims asserted against, such Contributor by reason
|
| 174 |
+
of your accepting any such warranty or additional liability.
|
| 175 |
+
|
| 176 |
+
END OF TERMS AND CONDITIONS
|
| 177 |
+
|
| 178 |
+
APPENDIX: How to apply the Apache License to your work.
|
| 179 |
+
|
| 180 |
+
To apply the Apache License to your work, attach the following
|
| 181 |
+
boilerplate notice, with the fields enclosed by brackets "[]"
|
| 182 |
+
replaced with your own identifying information. (Don't include
|
| 183 |
+
the brackets!) The text should be enclosed in the appropriate
|
| 184 |
+
comment syntax for the file format. We also recommend that a
|
| 185 |
+
file or class name and description of purpose be included on the
|
| 186 |
+
same "printed page" as the copyright notice for easier
|
| 187 |
+
identification within third-party archives.
|
| 188 |
+
|
| 189 |
+
Copyright 2026 BrainboxAI
|
| 190 |
+
|
| 191 |
+
Licensed under the Apache License, Version 2.0 (the "License");
|
| 192 |
+
you may not use this file except in compliance with the License.
|
| 193 |
+
You may obtain a copy of the License at
|
| 194 |
+
|
| 195 |
+
http://www.apache.org/licenses/LICENSE-2.0
|
| 196 |
+
|
| 197 |
+
Unless required by applicable law or agreed to in writing, software
|
| 198 |
+
distributed under the License is distributed on an "AS IS" BASIS,
|
| 199 |
+
WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
| 200 |
+
See the License for the specific language governing permissions and
|
| 201 |
+
limitations under the License.
|
NOTICE
ADDED
|
@@ -0,0 +1,29 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
Nitzotz (BrainboxAI/nitzotz)
|
| 2 |
+
Copyright 2026 BrainboxAI
|
| 3 |
+
|
| 4 |
+
This model is released under the Apache License, Version 2.0 (see LICENSE).
|
| 5 |
+
It is built from the following third-party work. Their licences and attribution requirements are kept here.
|
| 6 |
+
|
| 7 |
+
Encoder
|
| 8 |
+
- HalleluBERT-large (HalleluBERT/HalleluBERT_large), MIT License. The encoder weights of this model were initialised
|
| 9 |
+
from it and fine-tuned. Paper: arXiv 2510.21372.
|
| 10 |
+
|
| 11 |
+
Architecture, training method and runtime
|
| 12 |
+
- laya (github.com/NandhaKishorM/laya, NandhaKishorM / Convai Innovations), Apache License 2.0. The DecisionModel head
|
| 13 |
+
architecture, the option-marker scoring, the RLCD training objective and the laya runtime. The head of this model
|
| 14 |
+
was trained from scratch; no weights come from any laya checkpoint.
|
| 15 |
+
- ggmlc (github.com/monatis/ggmlc), used to compile the GGUF files. ggmlc/nitzotz_trunk.py and
|
| 16 |
+
ggmlc/compile_nitzotz.py are adapted from its examples/laya files (laya_trunk.py, compile_laya.py).
|
| 17 |
+
|
| 18 |
+
Training data
|
| 19 |
+
- HeQ, Hebrew Question Answering Dataset v1.1 (NNLP-IL, github.com/NNLP-IL/Hebrew-Question-Answering-Dataset, commit
|
| 20 |
+
f46d2ff), CC BY 4.0. Created by Webiks for MAFAT and the National NLP Program of Israel. Includes passages from
|
| 21 |
+
Hebrew Wikipedia and from Geektime, shared by the dataset authors under the dataset licence. Train split only.
|
| 22 |
+
Items were reformatted into yes/no and 4-option questions.
|
| 23 |
+
- MASSIVE (AmazonScience/massive, he-IL), CC BY 4.0, Amazon. FitzGerald et al., 2022. Train split only. Intent
|
| 24 |
+
labels were translated into short Hebrew labels.
|
| 25 |
+
- Synthetic Hebrew messages written by DeepSeek V4.1 Flash (deepseek-ai, MIT License) and labelled by DeepSeek V4.1
|
| 26 |
+
Flash and Gemma 4 31B (google/gemma-4-31b-it, Apache License 2.0), both served by DeepInfra through OpenRouter,
|
| 27 |
+
with no data retention. These synthetic data are not redistributed here.
|
| 28 |
+
|
| 29 |
+
No personal data and no private customer data were used.
|
README.md
ADDED
|
@@ -0,0 +1,478 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
language:
|
| 4 |
+
- he
|
| 5 |
+
library_name: laya
|
| 6 |
+
pipeline_tag: zero-shot-classification
|
| 7 |
+
base_model: HalleluBERT/HalleluBERT_large
|
| 8 |
+
tags:
|
| 9 |
+
- laya
|
| 10 |
+
- hebrew
|
| 11 |
+
- decision-model
|
| 12 |
+
- calibrated
|
| 13 |
+
- scam-detection
|
| 14 |
+
- routing
|
| 15 |
+
- gguf
|
| 16 |
+
datasets:
|
| 17 |
+
- AmazonScience/massive
|
| 18 |
+
---
|
| 19 |
+
|
| 20 |
+
> [!WARNING]
|
| 21 |
+
> **DRAFT, numbers will be replaced by the final model.** This card is a local draft. Nothing here is published, and the model is not on Hugging Face yet. The numbers are the current best results, from one checkpoint used for every test.
|
| 22 |
+
|
| 23 |
+

|
| 24 |
+
|
| 25 |
+

|
| 26 |
+
|
| 27 |
+
**What it is.** Nitzotz reads a Hebrew message and answers questions you type about it: pick one of several options,
|
| 28 |
+
give a score on a scale, or say yes or no to a claim. For every answer it gives a probability you can trust, so you
|
| 29 |
+
know when it is sure and when it is guessing. It does not write text, so it cannot make things up. It runs on a normal
|
| 30 |
+
laptop, with no internet connection and no cost per question.
|
| 31 |
+
|
| 32 |
+
**What it is for.** Deciding what to do with incoming messages: is this a scam, what kind of message is it, which
|
| 33 |
+
department should get it, how urgent is it. It is not a chatbot and it does not read long documents well
|
| 34 |
+
(see [Limitations](#limitations)).
|
| 35 |
+
|
| 36 |
+
**In numbers.** On 298 Hebrew messages it says correctly whether a message is a scam 82.5% of the
|
| 37 |
+
time. For context: 70% of those messages are not scams, so a model that always
|
| 38 |
+
says "not a scam" would score 70.5%; the two other open Hebrew-capable laya models score
|
| 39 |
+
32.2% (RoeiG) and 50.3% (laya-multilingual). The number that matters more is
|
| 40 |
+
the ranking: it gives real scams a higher probability than legitimate messages 88% of the time (AUC 0.88).
|
| 41 |
+
|
| 42 |
+
## Try it
|
| 43 |
+
|
| 44 |
+
**Live demo:** a Hugging Face Space will be linked here (not created yet).
|
| 45 |
+
|
| 46 |
+
**Weights will be published with the release.** The commands below show how it will be used; the files they name are not uploaded yet.
|
| 47 |
+
|
| 48 |
+
**Python** (the [laya](https://github.com/NandhaKishorM/laya) library, `pip install laya`):
|
| 49 |
+
|
| 50 |
+
```python
|
| 51 |
+
import laya
|
| 52 |
+
agent = laya.load("BrainboxAI/nitzotz")
|
| 53 |
+
q = {"scam": {"type": "noul", "instructions": "讛讛讜讚注讛 诪谞住讛 诇讙专讜诐 诇谞诪注谉 诇砖诇诐 讘诇讬 住讬讘讛 诇讙讬讟讬诪讬转."}}
|
| 54 |
+
print(agent.predict("讛讞讘讬诇讛 砖诇讱 诪注讜讻讘转. 诇砖讞专讜专 砖诇诐 12.90 讘拽讬砖讜专", q)["answers"]["scam"]["noul"])
|
| 55 |
+
```
|
| 56 |
+
|
| 57 |
+
**laya.exe** (the standalone binary from [ggmlc](https://github.com/monatis/ggmlc/releases), no Python needed; one
|
| 58 |
+
JSON request per line in, one answer per line out):
|
| 59 |
+
|
| 60 |
+
```bash
|
| 61 |
+
laya daemon nitzotz-q8_0.gguf --device vulkan
|
| 62 |
+
{"id":"1","state":"拽讜讚 讛讗讬诪讜转 砖诇讱 讛讜讗 000000","questions":{"scam":{"type":"noul","instructions":"讛讛讜讚注讛 讛讬讗 谞讬住讬讜谉 讛讜谞讗讛."}}}
|
| 63 |
+
```
|
| 64 |
+
|
| 65 |
+
The answer is the probability that the claim is true, for example `0.83`. Several questions in one call are answered
|
| 66 |
+
together in one pass.
|
| 67 |
+
|
| 68 |
+
## Benchmarks
|
| 69 |
+
|
| 70 |
+
8 frozen test sets, 3,990 questions in total, locked by checksum before any training data existed. All three
|
| 71 |
+
models answered exactly the same questions. We compare only the open laya models that read Hebrew:
|
| 72 |
+
|
| 73 |
+
- **Nitzotz**: this model.
|
| 74 |
+
- **[RoeiG/laya-hebrew](https://huggingface.co/RoeiG/laya-hebrew)**: a laya model on NeoDictaBERT. **Licence
|
| 75 |
+
CC-BY-NC-SA-4.0: non-commercial use only.** It is here as a reference point, you cannot use it in a product.
|
| 76 |
+
- **[laya-multilingual](https://huggingface.co/convaiinnovations/laya)** (convaiinnovations, the `multilingual`
|
| 77 |
+
subfolder): the original laya authors' model for 100+ languages. Apache-2.0.
|
| 78 |
+
|
| 79 |
+

|
| 80 |
+
|
| 81 |
+

|
| 82 |
+
|
| 83 |
+
**Full table.** Accuracy, the best in each row in bold. The last two columns say whether Nitzotz's difference from that
|
| 84 |
+
model is real or could be luck (a paired exact McNemar test on the same questions): "better" or "worse" means
|
| 85 |
+
p < 0.05, "tie" means the difference could be chance.
|
| 86 |
+
|
| 87 |
+
| Test (questions) | Nitzotz | RoeiG (non-commercial) | laya-multilingual | Chance | Nitzotz vs RoeiG | Nitzotz vs laya-multilingual |
|
| 88 |
+
|---|---|---|---|---|---|---|
|
| 89 |
+
| Scam or not? (298 messages) | **82.5%** | 32.2% | 50.3% | 50.0% | **better**<br>p<0.001 | **better**<br>p<0.001 |
|
| 90 |
+
| same, hard cases only (65) | **63.1%** | 32.3% | 35.4% | 50.0% | **better**<br>p<0.001 | **better**<br>p=0.008 |
|
| 91 |
+
| Message type, 6 options (298) | **75.5%** | 52.7% | 20.8% | 16.7% | **better**<br>p<0.001 | **better**<br>p<0.001 |
|
| 92 |
+
| same, hard cases only (65) | **53.8%** | 44.6% | 24.6% | 16.7% | tie<br>p=0.34 | **better**<br>p<0.001 |
|
| 93 |
+
| Support ticket type, 5 options (30) | **83.3%** | 80.0% | 56.7% | 20.0% | tie<br>p=1.00 | tie<br>p=0.06 |
|
| 94 |
+
| Ticket urgency, 5 levels (30) | **60.0%** | 36.7% | 33.3% | 20.0% | tie<br>p=0.17 | tie<br>p=0.12 |
|
| 95 |
+
| Paying customer? yes/no (30) | **63.3%** | 56.7% | 50.0% | 50.0% | tie<br>p=0.50 | tie<br>p=0.54 |
|
| 96 |
+
| Voice command intent, 20 options (500) | **88.8%** | 72.6% | 47.4% | 5.0% | **better**<br>p<0.001 | **better**<br>p<0.001 |
|
| 97 |
+
| Voice command intent, 4 options (500) | **96.2%** | 90.8% | 69.4% | 25.0% | **better**<br>p<0.001 | **better**<br>p<0.001 |
|
| 98 |
+
| News topic, 7 options (204) | 79.4% | **82.3%** | 66.2% | 14.3% | tie<br>p=0.42 | **better**<br>p<0.001 |
|
| 99 |
+
| Does the passage support this answer? (600) | 93.8% | **94.2%** | 48.8% | 50.0% | tie<br>p=0.90 | **better**<br>p<0.001 |
|
| 100 |
+
| Plausible answer the passage does not give (600) | **91.0%** | 53.0% | 53.7% | 50.0% | **better**<br>p<0.001 | **better**<br>p<0.001 |
|
| 101 |
+
| Reading comprehension, 4 options (900) | 52.7% | **75.4%** | 31.4% | 25.0% | **worse**<br>p<0.001 | **better**<br>p<0.001 |
|
| 102 |
+
|
| 103 |
+
How to read it:
|
| 104 |
+
- **MASSIVE** (voice commands): Nitzotz was trained on MASSIVE's training commands. The test commands are different,
|
| 105 |
+
but written by the same people in the same style, so this test is easier for Nitzotz than for the others.
|
| 106 |
+
- **HeQ**: both Nitzotz (for these two rows) and RoeiG were trained on HeQ's training passages. The test passages are
|
| 107 |
+
different ones.
|
| 108 |
+
- **The support-ticket rows have only 30 questions.** No difference there is reliable.
|
| 109 |
+
- **The spam and scam test has a lean**: 70% of its messages are not scams. The
|
| 110 |
+
"chance" line (50%) is a coin flip, not the best blind strategy.
|
| 111 |
+
|
| 112 |
+
### Where RoeiG leads
|
| 113 |
+
|
| 114 |
+
We say this plainly because you should know it before you choose:
|
| 115 |
+
|
| 116 |
+
- **News topic, 7 options (204)**: RoeiG 82.3%, Nitzotz 79.4%. No clear difference (p=0.42).
|
| 117 |
+
- **Does the passage support this answer? (600)**: RoeiG 94.2%, Nitzotz 93.8%. No clear difference (p=0.90).
|
| 118 |
+
- **Reading comprehension, 4 options (900)**: RoeiG 75.4%, Nitzotz 52.7%. Nitzotz worse (p<0.001).
|
| 119 |
+
|
| 120 |
+
**Reading a passage and answering questions about it (Belebele) is RoeiG's strength and Nitzotz's weakness.** RoeiG
|
| 121 |
+
was trained on reading-comprehension data; Nitzotz was not built for it. If your task is reading documents, Nitzotz is
|
| 122 |
+
the wrong model. Remember that RoeiG cannot be used commercially.
|
| 123 |
+
|
| 124 |
+
## Why you can trust it
|
| 125 |
+
|
| 126 |
+

|
| 127 |
+
|
| 128 |
+
**The probabilities mean something.** Take all the answers where Nitzotz said it was about 70 to 80% sure, and count
|
| 129 |
+
how many were right: the chart above does that for every confidence level, over 3,990 test questions.
|
| 130 |
+
Where the dots sit above the line, Nitzotz is more often right than it claims (it is modest); where they sit below, it
|
| 131 |
+
is over-confident. This is what lets you set thresholds (see the next section). The per-question-type temperatures
|
| 132 |
+
were fitted on held-out training items, never on the test sets.
|
| 133 |
+
|
| 134 |
+
**It reads the text.** With the message removed and only the question left, its accuracy falls to
|
| 135 |
+
70.5% on the scam question and 7.7% on message type. So the
|
| 136 |
+
answers come from the message, not from the wording of the question.
|
| 137 |
+
|
| 138 |
+

|
| 139 |
+
|
| 140 |
+
**It is fast on ordinary hardware.** On one laptop (Intel Core Ultra 9 285H laptop, built-in Arc 140T GPU, Windows 11), one question at a time:
|
| 141 |
+
the Q8 file answers in a median of 140 ms (slowest 10%: 187 ms or more), the F16 file in 154 ms.
|
| 142 |
+
Through Python on the same GPU Nitzotz takes 294 ms and RoeiG 1005 ms (measured on the same machine, but
|
| 143 |
+
in separate runs on 50 and 40 test questions of similar length, about 134 and 136 tokens). laya-multilingual was
|
| 144 |
+
not measured on this machine, so it has no speed on the chart. The Q8 file gives the same top answer as the full
|
| 145 |
+
Python model on 50/50 check questions.
|
| 146 |
+
|
| 147 |
+
## Use it in your business
|
| 148 |
+
|
| 149 |
+

|
| 150 |
+
|
| 151 |
+
**The idea ("Ramzor", traffic light).** Every incoming message gets one or more quick questions. The probability
|
| 152 |
+
decides what happens next:
|
| 153 |
+
|
| 154 |
+
- **Green** (confident it is fine): handle automatically, file it, tag it, route it.
|
| 155 |
+
- **Yellow** (not sure): send it to a person, or to a large language model if you use one.
|
| 156 |
+
- **Red** (confident it is a scam): block or quarantine it.
|
| 157 |
+
|
| 158 |
+
The drawing is a worked example on the 298 test messages. The upper threshold, 0.52, is the one chosen for the
|
| 159 |
+
scam question on held-out training messages (never on the test); the lower one, 0.35, was picked by hand. With them,
|
| 160 |
+
183 messages go to green (16 of them are in fact scams), 48 to yellow (21 scams), and 67 to red (16 of them are
|
| 161 |
+
in fact legitimate). So today most messages still need a look, and red should mean "quarantine and check", not
|
| 162 |
+
"delete". Pick your own thresholds on a sample of your own messages, and decide how many mistakes in green and red
|
| 163 |
+
you can live with. At the 0.52 threshold alone, the scam answer is right 82.2% of the time overall and
|
| 164 |
+
64.6% on the 65 hard cases.
|
| 165 |
+
|
| 166 |
+
The model is cheap enough to run on every message. The person (or the LLM) only sees the yellow part. **Do not use it
|
| 167 |
+
as the only line of defence for decisions that can hurt someone.**
|
| 168 |
+
|
| 169 |
+
## Training data and transparency
|
| 170 |
+
|
| 171 |
+
58,000 training items, one pass over the data, on a rented GPU.
|
| 172 |
+
|
| 173 |
+
| Source | Items | Licence | How it was made |
|
| 174 |
+
|---|---|---|---|
|
| 175 |
+
| Synthetic: Message type, 6 classes | 14,000 | teacher output, project-owned (DeepSeek MIT; Gemma Apache-2.0) | written by DeepSeek V4.1 Flash, labelled independently by both models |
|
| 176 |
+
| Synthetic: Routing to a department (3 to 6 options) | 8,750 | teacher output, project-owned (DeepSeek MIT; Gemma Apache-2.0) | written by DeepSeek V4.1 Flash, labelled independently by both models |
|
| 177 |
+
| Synthetic: Yes/no claims about a message | 8,750 | teacher output, project-owned (DeepSeek MIT; Gemma Apache-2.0) | written by DeepSeek V4.1 Flash, labelled independently by both models |
|
| 178 |
+
| Synthetic: Urgency, 5 levels | 3,500 | teacher output, project-owned (DeepSeek MIT; Gemma Apache-2.0) | written by DeepSeek V4.1 Flash, labelled independently by both models |
|
| 179 |
+
| HeQ: is the proposed answer supported (yes/no) | 6,000 | CC BY 4.0 | the dataset's gold labels, train split only |
|
| 180 |
+
| HeQ: plausible answer to an unanswerable question | 5,000 | CC BY 4.0 | the dataset's gold labels, train split only |
|
| 181 |
+
| HeQ: 4-option reading | 5,000 | CC BY 4.0 | the dataset's gold labels, train split only |
|
| 182 |
+
| MASSIVE he-IL: voice command intent | 7,000 | CC BY 4.0 | the dataset's gold labels, train split only |
|
| 183 |
+
|
| 184 |
+
- **Synthetic messages (35,000 items).** DeepSeek V4.1 Flash (MIT) wrote Israeli-style SMS, WhatsApp and email messages
|
| 185 |
+
from a plan (intended label, topic, tone, varied fake phone numbers and links). DeepSeek and Gemma 4 31B
|
| 186 |
+
(Apache-2.0) then each labelled every message on their own. A message was kept only if both agreed; the two models
|
| 187 |
+
agreed on 94% of them, and 93.7% of the 37,429 written messages were kept. Both ran
|
| 188 |
+
through DeepInfra (via OpenRouter), with data retention off and "thinking" mode off. Cost: 5.39 USD. The
|
| 189 |
+
synthetic data is not published.
|
| 190 |
+
- **Open data.** HeQ v1.1 (CC BY 4.0; about half of its questions are on Geektime articles, shared by the HeQ authors under
|
| 191 |
+
the same licence) and MASSIVE he-IL (CC BY 4.0). Training splits only.
|
| 192 |
+
- **No leaks from the tests.** Every training item was compared with every test question; anything sharing an
|
| 193 |
+
8-word run with a test text was dropped, and so were MASSIVE commands equal to a test command.
|
| 194 |
+
- **Not used:** no output of closed commercial chatbots, no DICTA model, no non-commercial or share-alike data.
|
| 195 |
+
- **Encoder:** HalleluBERT-large (MIT), 357.1M parameters, fine-tuned; laya head 26.5M parameters, trained from scratch.
|
| 196 |
+
- **The reading encoder is merged into this checkpoint.** HalleluBERT was first trained to find answers in HeQ passages (27,085 HeQ training questions, CC BY 4.0), then given a laya head and fine-tuned on the same training set as every other row. The two HeQ rows use the same checkpoint as the rest of the table, not a separate one.
|
| 197 |
+
|
| 198 |
+
**About the tests.**
|
| 199 |
+
|
| 200 |
+
| Test | Questions | Source and licence |
|
| 201 |
+
|---|---|---|
|
| 202 |
+
| Spam and business messages | 596 (298 messages, 2 questions each) | in-house, Israeli SMS, WhatsApp and email style, 65 hard cases |
|
| 203 |
+
| Support tickets (triage30) | 90 (30 tickets, 3 questions each) | in-house |
|
| 204 |
+
| MASSIVE he-IL, 20 and 4 options | 500 + 500 | MASSIVE test split, CC BY 4.0 |
|
| 205 |
+
| SIB-200 news topic | 204 | CC BY-SA 4.0, used for testing only |
|
| 206 |
+
| Belebele reading | 900 | CC BY-SA 4.0, used for testing only |
|
| 207 |
+
| HeQ verify and unanswerable | 600 + 600 | HeQ v1.1 test split, CC BY 4.0 |
|
| 208 |
+
|
| 209 |
+
Caveats that change how much to trust the numbers:
|
| 210 |
+
- **The spam and business test was written by an AI model and checked by an AI model, not by a person.** Real inboxes
|
| 211 |
+
will look different. It was frozen after that review (4 labels changed, 2 messages removed).
|
| 212 |
+
- **One training run, one random seed.** Differences of 1 to 3 points between models may be run-to-run noise.
|
| 213 |
+
- **The training labels come from two AI models.** Where both are wrong in the same way, Nitzotz learned their mistake.
|
| 214 |
+
- HeQ's wrong answers in the test were picked by code, not checked by a person.
|
| 215 |
+
|
| 216 |
+
## Limitations
|
| 217 |
+
|
| 218 |
+
- **It does not read documents well.** Belebele reading comprehension: 52.7%, where a blind guess gets 25%.
|
| 219 |
+
Do not ask it whether a long document supports a claim.
|
| 220 |
+
- **HeQ answer checking looks strong, with a catch.** 93.8% with the passage, but still 50.0% with the
|
| 221 |
+
passage removed (RoeiG: 80.0%). A lot of that score comes from the question and answer alone, not from reading.
|
| 222 |
+
- **Numbers, dates, amounts and rules**: not trained and not measured. Compute them in code and pass the result in.
|
| 223 |
+
- **Hard cases are the weak spot**: scams written to look legitimate (a "supplier" changing bank details, the "CEO"
|
| 224 |
+
asking for a transfer) and real messages that look like scams (a real bank alert with a link, a real verification
|
| 225 |
+
code). On the 65 hard cases the scam question is right 63.1% of the time.
|
| 226 |
+
- **Urgency is subjective.** Even the two teacher models matched the intended urgency only about 62 to 64% of the time.
|
| 227 |
+
- **Sarcasm and irony** are probably read literally. Not measured.
|
| 228 |
+
- **512 tokens** (roughly 300 to 400 Hebrew words) per question. A longer message is cut from the end without a warning.
|
| 229 |
+
- **Hebrew only.** Not trained or tested on English or Arabic.
|
| 230 |
+
- **Not a safety system on its own.** It makes mistakes in both directions. Keep a person in the loop for anything
|
| 231 |
+
that can hurt someone.
|
| 232 |
+
|
| 233 |
+
## Licence and attribution
|
| 234 |
+
|
| 235 |
+
**Apache-2.0** for the weights, the GGUF files and the code. Commercial use is allowed. Built on:
|
| 236 |
+
|
| 237 |
+
- [HalleluBERT-large](https://huggingface.co/HalleluBERT/HalleluBERT_large): the encoder, MIT.
|
| 238 |
+
- [laya](https://github.com/NandhaKishorM/laya) (NandhaKishorM, Convai Innovations): architecture, training method and
|
| 239 |
+
runtime, Apache-2.0. No laya weights are used.
|
| 240 |
+
- [HeQ](https://github.com/NNLP-IL/Hebrew-Question-Answering-Dataset): CC BY 4.0, by Webiks for MAFAT and the Israeli
|
| 241 |
+
National NLP Program (NNLP-IL); includes Geektime passages.
|
| 242 |
+
- [MASSIVE](https://huggingface.co/datasets/AmazonScience/massive): CC BY 4.0, Amazon (FitzGerald et al., 2022).
|
| 243 |
+
- **DeepSeek V4.1 Flash** (MIT) and **Gemma 4 31B** (Apache-2.0), used through DeepInfra as data writer and labellers.
|
| 244 |
+
- [ggmlc](https://github.com/monatis/ggmlc) for the GGUF files.
|
| 245 |
+
- Belebele and SIB-200 (CC BY-SA 4.0) were used only to test, never to train.
|
| 246 |
+
|
| 247 |
+
The full notice is in `NOTICE`.
|
| 248 |
+
|
| 249 |
+
---
|
| 250 |
+
|
| 251 |
+
<div dir="rtl">
|
| 252 |
+
|
| 253 |
+
## 谞讬爪讜抓, 讘注讘专讬转
|
| 254 |
+
|
| 255 |
+
> [!WARNING]
|
| 256 |
+
> **讟讬讜讟讛, 讛诪住驻专讬诐 讬讜讞诇驻讜 讘诪讜讚诇 讛住讜驻讬.** 讝讛 讻专讟讬住 讟讬讜讟讛 诪拽讜诪讬. 砖讜诐 讚讘专 讻讗谉 诇讗 驻讜专住诐, 讜讛诪讜讚诇 注讜讚 诇讗 谞诪爪讗 讘-Hugging Face. 讛诪住驻专讬诐 讛诐 讛转讜爪讗讜转 讛讟讜讘讜转 讘讬讜转专 讻专讙注, 诪讙专住讛 讗讞转 砖谞讘讚拽讛 讘讻诇 讛诪讘讞谞讬诐.
|
| 257 |
+
|
| 258 |
+

|
| 259 |
+
|
| 260 |
+

|
| 261 |
+
|
| 262 |
+
**诪讛 讝讛.** 谞讬爪讜抓 拽讜专讗 讛讜讚注讛 讘注讘专讬转 讜注讜谞讛 注诇 砖讗诇讜转 砖讗转诐 诪拽诇讬讚讬诐 注诇讬讛: 诇讘讞讜专 讗讞转 诪讻诪讛 讗驻砖专讜讬讜转, 诇转转 爪讬讜谉 讘住讜诇诐, 讗讜
|
| 263 |
+
诇注谞讜转 讻谉 讗讜 诇讗 注诇 讟注谞讛. 注诇 讻诇 转砖讜讘讛 讛讜讗 谞讜转谉 讛住转讘专讜转 砖讗驻砖专 诇住诪讜讱 注诇讬讛, 讻讱 砖讬讜讚注讬诐 诪转讬 讛讜讗 讘讟讜讞 讜诪转讬 讛讜讗 诪谞讞砖. 讛讜讗 诇讗
|
| 264 |
+
讻讜转讘 讟拽住讟, 讜诇讻谉 讛讜讗 诇讗 讬讻讜诇 诇讛诪爪讬讗 讚讘专讬诐. 讛讜讗 专抓 注诇 诪讞砖讘 谞讬讬讚 专讙讬诇, 讘诇讬 讗讬谞讟专谞讟 讜讘诇讬 转砖诇讜诐 注诇 讻诇 砖讗诇讛.
|
| 265 |
+
|
| 266 |
+
**讘砖讘讬诇 诪讛.** 诇讛讞诇讬讟 诪讛 注讜砖讬诐 注诐 讛讜讚注讜转 谞讻谞住讜转: 讛讗诐 讝讜 讛讜谞讗讛, 讗讬讝讛 住讜讙 讛讜讚注讛 讝讜, 诇讗讬讝讜 诪讞诇拽讛 诇讛注讘讬专, 讻诪讛 讝讛 讚讞讜祝. 讝讛
|
| 267 |
+
诇讗 爪'讗讟讘讜讟, 讜讛讜讗 诇讗 讟讜讘 讘拽专讬讗转 诪住诪讻讬诐 讗专讜讻讬诐 (专讗讜 [诪讙讘诇讜转](#诪讙讘诇讜转)).
|
| 268 |
+
|
| 269 |
+
**讘诪住驻专讬诐.** 注诇 298 讛讜讚注讜转 讘注讘专讬转 讛讜讗 拽讜讘注 谞讻讜谉 讗诐 讛讛讜讚注讛 讛讬讗 讛讜谞讗讛 讘-82.5% 诪讛诪拽专讬诐. 讘砖讘讬诇
|
| 270 |
+
驻专讜驻讜专爪讬讛: 70% 诪讛讛讜讚注讜转 讛讗诇讛 讛谉 诇讗 讛讜谞讗讛, 讻讱 砖诪讜讚诇 砖转诪讬讚 注讜谞讛 "诇讗 讛讜谞讗讛" 讛讬讛 诪拽讘诇 70.5%.
|
| 271 |
+
砖谞讬 诪讜讚诇讬 laya 讛驻转讜讞讬诐 讛讗讞专讬诐 砖拽讜专讗讬诐 注讘专讬转 诪拽讘诇讬诐 32.2% (RoeiG) 讜-50.3% (laya-multilingual). 讛诪住驻专
|
| 272 |
+
砖讞砖讜讘 讬讜转专 讛讜讗 讛讚讬专讜讙: 讛讜讗 谞讜转谉 诇讛讜谞讗讛 讗诪讬转讬转 讛住转讘专讜转 讙讘讜讛讛 讬讜转专 诪讗砖专 诇讛讜讚注讛 转拽讬谞讛 讘-88% 诪讛诪拽专讬诐 (AUC 0.88).
|
| 273 |
+
|
| 274 |
+
### 诇谞住讜转
|
| 275 |
+
|
| 276 |
+
**讛讚讙诪讛 讞讬讛:** 讻讗谉 讬讜驻讬注 拽讬砖讜专 诇-Space 讘-Hugging Face (注讜讚 诇讗 谞讜爪专).
|
| 277 |
+
|
| 278 |
+
**讛诪砖拽讜诇讜转 讬驻讜专住诪讜 讬讞讚 注诐 讛讛砖拽讛.** 讛驻拽讜讚讜转 诇诪讟讛 诪专讗讜转 讗讬讱 讝讛 讬讬专讗讛, 讛拽讘爪讬诐 砖讛谉 诪讝讻讬专讜转 注讜讚 诇讗 讛讜注诇讜.
|
| 279 |
+
|
| 280 |
+
**讘驻讬讬转讜谉** (讛住驻专讬讬讛 [laya](https://github.com/NandhaKishorM/laya), `pip install laya`):
|
| 281 |
+
|
| 282 |
+
</div>
|
| 283 |
+
|
| 284 |
+
```python
|
| 285 |
+
import laya
|
| 286 |
+
agent = laya.load("BrainboxAI/nitzotz")
|
| 287 |
+
q = {"scam": {"type": "noul", "instructions": "讛讛讜讚注讛 诪谞住讛 诇讙专讜诐 诇谞诪注谉 诇砖诇诐 讘诇讬 住讬讘讛 诇讙讬讟讬诪讬转."}}
|
| 288 |
+
print(agent.predict("讛讞讘讬诇讛 砖诇讱 诪注讜讻讘转. 诇砖讞专讜专 砖诇诐 12.90 讘拽讬砖讜专", q)["answers"]["scam"]["noul"])
|
| 289 |
+
```
|
| 290 |
+
|
| 291 |
+
<div dir="rtl">
|
| 292 |
+
|
| 293 |
+
**讘诇讬 驻讬讬转讜谉, 注诐 laya.exe** (转讜讻谞讛 注爪诪讗讬转 诪-[ggmlc](https://github.com/monatis/ggmlc/releases). 讘拽砖转 JSON 讗讞转 讘讻诇
|
| 294 |
+
砖讜专讛 谞讻谞住转, 转砖讜讘讛 讗讞转 讘讻诇 砖讜专讛 讬讜爪讗转):
|
| 295 |
+
|
| 296 |
+
</div>
|
| 297 |
+
|
| 298 |
+
```bash
|
| 299 |
+
laya daemon nitzotz-q8_0.gguf --device vulkan
|
| 300 |
+
{"id":"1","state":"拽讜讚 讛讗讬诪讜转 砖诇讱 讛讜讗 000000","questions":{"scam":{"type":"noul","instructions":"讛讛讜讚注讛 讛讬讗 谞讬住讬讜谉 讛讜谞讗讛."}}}
|
| 301 |
+
```
|
| 302 |
+
|
| 303 |
+
<div dir="rtl">
|
| 304 |
+
|
| 305 |
+
讛转砖讜讘讛 讛讬讗 讛讛住转讘专讜转 砖讛讟注谞讛 谞讻讜谞讛, 诇诪砖诇 `0.83`. 讻诪讛 砖讗诇讜转 讘拽专讬讗讛 讗讞转 谞注谞讜转 讬讞讚, 讘诪注讘专 讗讞讚.
|
| 306 |
+
|
| 307 |
+
### 诪讘讞谞讬诐
|
| 308 |
+
|
| 309 |
+
8 住讟讬诐 砖诇 诪讘讞谉, 3,990 砖讗诇讜转 讘住讱 讛讻讜诇, 砖谞谞注诇讜 讘讟讘讬注转 讗爪讘注 诇驻谞讬 砖谞讜爪专 驻专讬讟 讗讬诪讜谉 讗讞讚. 砖诇讜砖转 讛诪讜讚诇讬诐 注谞讜 注诇
|
| 310 |
+
讗讜转谉 砖讗诇讜转 讘讚讬讜拽. 诪砖讜讜讬诐 专拽 诇诪讜讚诇讬 laya 驻转讜讞讬诐 砖拽讜专讗讬诐 注讘专讬转:
|
| 311 |
+
|
| 312 |
+
- **谞讬爪讜抓**: 讛诪讜讚诇 讛讝讛.
|
| 313 |
+
- **[RoeiG/laya-hebrew](https://huggingface.co/RoeiG/laya-hebrew)**: 诪讜讚诇 laya 注诇 NeoDictaBERT. **专讬砖讬讜谉
|
| 314 |
+
CC-BY-NC-SA-4.0: 诇砖讬诪讜砖 诇讗 诪住讞专讬 讘诇讘讚.** 讛讜讗 讻讗谉 讻谞拽讜讚转 讛砖讜讜讗讛. 讗讬 讗驻砖专 诇讛砖转诪砖 讘讜 讘诪讜爪专.
|
| 315 |
+
- **[laya-multilingual](https://huggingface.co/convaiinnovations/laya)** (convaiinnovations, 转讬拽讬讬转 `multilingual`):
|
| 316 |
+
讛诪讜讚诇 砖诇 讬讜爪专讬 laya 诇讬讜转专 诪-100 砖驻讜转. Apache-2.0.
|
| 317 |
+
|
| 318 |
+

|
| 319 |
+
|
| 320 |
+

|
| 321 |
+
|
| 322 |
+
**讛讟讘诇讛 讛诪诇讗讛.** 讗讞讜讝 讛转砖讜讘讜转 讛谞讻讜谞讜转, 讛讟讜讘 讘讬讜转专 讘讻诇 砖讜专讛 诪讜讚讙砖. 砖转讬 讛注诪讜讚讜转 讛讗讞专讜谞讜转 讗讜诪专讜转 讗诐 讛讛讘讚诇 砖诇 谞讬爪讜抓
|
| 323 |
+
诪讛诪讜讚诇 讛讝讛 讗诪讬转讬 讗讜 砖讗讜诇讬 讝讛 诪讝诇 (诪讘讞谉 McNemar 诪讚讜讬拽 注诇 讗讜转谉 砖讗诇讜转 讘讚讬讜拽): "讟讜讘 讬讜转专" 讗讜 "讞诇砖 讬讜转专" 驻讬专讜砖讜 p
|
| 324 |
+
拽讟谉 诪-0.05, "转讬拽讜" 驻讬专讜砖讜 砖讛讛讘讚诇 讬讻讜诇 诇讛讬讜转 诪拽专讬.
|
| 325 |
+
|
| 326 |
+
| 诪讘讞谉 (诪住驻专 砖讗诇讜转) | 谞讬爪讜抓 | RoeiG (诇讗 诪住讞专讬) | laya-multilingual | 谞讬讞讜砖 | 谞讬爪讜抓 诪讜诇 RoeiG | 谞讬爪讜抓 诪讜诇 laya-multilingual |
|
| 327 |
+
|---|---|---|---|---|---|---|
|
| 328 |
+
| 讛讜谞讗讛 讗讜 诇讗? (298 讛讜讚注讜转) | **82.5%** | 32.2% | 50.3% | 50.0% | **讟讜讘 讬讜转专**<br>p<0.001 | **讟讜讘 讬讜转专**<br>p<0.001 |
|
| 329 |
+
| 讗讜转讜 讚讘专, 专拽 讛诪拽专讬诐 讛拽砖讬诐 (65) | **63.1%** | 32.3% | 35.4% | 50.0% | **讟讜讘 讬讜转专**<br>p<0.001 | **讟讜讘 讬讜转专**<br>p=0.008 |
|
| 330 |
+
| 住讜讙 讛讛讜讚注讛, 6 讗驻砖专讜讬讜转 (298) | **75.5%** | 52.7% | 20.8% | 16.7% | **讟讜讘 讬讜转专**<br>p<0.001 | **讟讜讘 讬讜转专**<br>p<0.001 |
|
| 331 |
+
| 讗讜转讜 讚讘专, 专拽 讛诪拽专讬诐 讛拽砖讬诐 (65) | **53.8%** | 44.6% | 24.6% | 16.7% | 转讬拽讜<br>p=0.34 | **讟讜讘 讬讜转专**<br>p<0.001 |
|
| 332 |
+
| 住讜讙 驻谞讬讬转 转诪讬讻讛, 5 讗驻砖专讜讬讜转 (30) | **83.3%** | 80.0% | 56.7% | 20.0% | 转讬拽讜<br>p=1.00 | 转讬拽讜<br>p=0.06 |
|
| 333 |
+
| 讚讞讬驻讜转 讛驻谞讬讬讛, 5 专诪讜转 (30) | **60.0%** | 36.7% | 33.3% | 20.0% | 转讬拽讜<br>p=0.17 | 转讬拽讜<br>p=0.12 |
|
| 334 |
+
| 诇拽讜讞 诪砖诇诐? 讻谉/诇讗 (30) | **63.3%** | 56.7% | 50.0% | 50.0% | 转讬拽讜<br>p=0.50 | 转讬拽讜<br>p=0.54 |
|
| 335 |
+
| 讻讜讜谞转 驻拽讜讚讛 拽讜诇讬转, 20 讗驻砖专讜讬讜转 (500) | **88.8%** | 72.6% | 47.4% | 5.0% | **讟讜讘 讬讜转专**<br>p<0.001 | **讟讜讘 讬讜转专**<br>p<0.001 |
|
| 336 |
+
| 讻讜讜谞转 驻拽讜讚讛 拽讜诇讬转, 4 讗驻砖专讜讬讜转 (500) | **96.2%** | 90.8% | 69.4% | 25.0% | **讟讜讘 讬讜转专**<br>p<0.001 | **讟讜讘 讬讜转专**<br>p<0.001 |
|
| 337 |
+
| 谞讜砖讗 砖诇 讬讚讬注讛, 7 讗驻砖专讜讬讜转 (204) | 79.4% | **82.3%** | 66.2% | 14.3% | 转讬拽讜<br>p=0.42 | **讟讜讘 讬讜转专**<br>p<0.001 |
|
| 338 |
+
| 讛讗诐 讛拽讟注 转讜诪讱 讘转砖讜讘讛? (600) | 93.8% | **94.2%** | 48.8% | 50.0% | 转讬拽讜<br>p=0.90 | **讟讜讘 讬讜转专**<br>p<0.001 |
|
| 339 |
+
| 转砖讜讘讛 住讘讬专讛 砖讛拽讟注 诇讗 谞讜转谉 (600) | **91.0%** | 53.0% | 53.7% | 50.0% | **讟讜讘 讬讜转专**<br>p<0.001 | **讟讜讘 讬讜转专**<br>p<0.001 |
|
| 340 |
+
| 讛讘谞转 讛谞拽专讗, 4 讗驻砖专讜讬讜转 (900) | 52.7% | **75.4%** | 31.4% | 25.0% | **讞诇砖 讬讜转专**<br>p<0.001 | **讟讜讘 讬讜转专**<br>p<0.001 |
|
| 341 |
+
|
| 342 |
+
讗讬讱 诇拽专讜讗 讗转 讝讛:
|
| 343 |
+
- **MASSIVE** (驻拽讜讚讜转 拽讜诇讬讜转): 谞讬爪讜抓 讗讜诪谉 注诇 驻拽讜讚讜转 讛讗讬诪讜谉 砖诇 MASSIVE. 驻拽讜讚讜转 讛诪讘讞谉 讗讞专讜转, 讗讘诇 谞讻转讘讜 讘讬讚讬 讗讜转诐
|
| 344 |
+
讗谞砖讬诐 讜讘讗讜转讜 住讙谞讜谉, 讜诇讻谉 讛诪讘讞谉 讛讝讛 拽诇 讬讜转专 诇谞讬爪讜抓 诪讗砖专 诇讗讞专讬诐.
|
| 345 |
+
- **HeQ**: 讙诐 谞讬爪讜抓 (讘砖转讬 讛砖讜专讜转 讛讗诇讛) 讜讙诐 RoeiG 讗讜诪谞讜 注诇 拽讟注讬 讛讗讬诪讜谉 砖诇 HeQ. 拽讟注讬 讛诪讘讞谉 讗讞专讬诐.
|
| 346 |
+
- **讘砖讜专讜转 砖诇 驻谞讬讜转 讛转诪讬讻讛 讬砖 专拽 30 砖讗诇讜转.** 砖讜诐 讛讘讚诇 砖诐 诇讗 讗诪讬谉.
|
| 347 |
+
- **诪讘讞谉 讛住驻讗诐 讜讛讛讜谞讗讜转 诇讗 诪讗讜讝谉**: 70% 诪讛讛讜讚注讜转 讘讜 讛谉 诇讗 讛讜谞讗讛. 拽讜 讛"谞讬讞讜砖" (50%) 讛讜讗 讛讟诇转
|
| 348 |
+
诪讟讘注, 诇讗 讛讗住讟专讟讙讬讛 讛注讬讜讜专转 讛讟讜讘讛 讘讬讜转专.
|
| 349 |
+
|
| 350 |
+
### 讗讬驻讛 RoeiG 诪讜讘讬诇
|
| 351 |
+
|
| 352 |
+
讗讜诪专讬诐 讗转 讝讛 讘驻砖讟讜转, 讻讬 讻讚讗讬 诇讚注转 讗转 讝讛 诇驻谞讬 砖讘讜讞专讬诐:
|
| 353 |
+
|
| 354 |
+
- **谞讜砖讗 砖诇 讬讚讬注讛, 7 讗驻砖专讜讬讜转 (204)**: RoeiG 82.3%, 谞讬爪讜抓 79.4%. 讗讬谉 讛讘讚诇 讘专讜专 (p=0.42).
|
| 355 |
+
- **讛讗诐 讛拽讟注 转讜诪讱 讘转砖讜讘讛? (600)**: RoeiG 94.2%, 谞讬爪讜抓 93.8%. 讗讬谉 讛讘讚诇 讘专讜专 (p=0.90).
|
| 356 |
+
- **讛讘谞转 讛谞拽专讗, 4 讗驻砖专讜讬讜转 (900)**: RoeiG 75.4%, 谞讬爪讜抓 52.7%. 谞讬爪讜抓 讞诇砖 讬讜转专 (p<0.001).
|
| 357 |
+
|
| 358 |
+
**诇拽专讜讗 拽讟注 讜诇注谞讜转 注诇讬讜 砖讗诇讜转 (Belebele) 讝讛 讛讞讜讝拽 砖诇 RoeiG 讜讛讞讜诇砖讛 砖诇 谞讬爪讜抓.** RoeiG 讗讜诪谉 注诇 谞转讜谞讬 讛讘谞转 讛谞拽专讗,
|
| 359 |
+
讜谞讬爪讜抓 诇讗 谞讘谞讛 诇讝讛. 讗诐 讛诪砖讬诪讛 砖诇讻诐 讛讬讗 拽专讬讗转 诪住诪讻讬诐, 谞讬爪讜抓 讛讜讗 讛诪讜讚诇 讛诇讗 谞讻讜谉. 讜讝讻专讜 砖-RoeiG 讗住讜专 诇砖讬诪讜砖 诪住讞专讬.
|
| 360 |
+
|
| 361 |
+
### 诇诪讛 讗驻砖专 诇住诪讜讱 注诇讬讜
|
| 362 |
+
|
| 363 |
+

|
| 364 |
+
|
| 365 |
+
**诇讛住转讘专讜讬讜转 讬砖 诪砖诪注讜转.** 拽讞讜 讗转 讻诇 讛转砖讜讘讜转 砖讘讛谉 谞讬爪讜抓 讗诪锟斤拷 砖讛讜讗 讘讟讜讞 讘注专讱 讘-70 注讚 80%, 讜住驻专讜 讻诪讛 诪讛谉 讛讬讜 谞讻讜谞讜转.
|
| 366 |
+
讛讙专祝 注讜砖讛 讗转 讝讛 诇讻诇 专诪转 讘讬讟讞讜谉, 注诇 3,990 砖讗诇讜转 诪讘讞谉. 讻砖讛谞拽讜讚讜转 诪注诇 讛拽讜, 谞讬爪讜抓 爪讜讚拽 讬讜转专 诪诪讛 砖讛讜讗 讗讜诪专 (讛讜讗
|
| 367 |
+
爪谞讜注). 讻砖讛谉 诪转讞转, 讛讜讗 讘讟讜讞 讘注爪诪讜 讬讜转专 诪讚讬. 讝讛 诪讛 砖诪讗驻砖专 诇拽讘讜注 住驻讬诐 (讘驻专拽 讛讘讗). 讛讻讬讜诇 谞注砖讛 注诇 驻专讬讟讬 讗讬诪讜谉 砖讛讜驻专讚讜
|
| 368 |
+
诪专讗砖, 讗祝 驻注诐 诇讗 注诇 讛诪讘讞谉.
|
| 369 |
+
|
| 370 |
+
**讛讜讗 讘讗诪转 拽讜专讗 讗转 讛讟拽住讟.** 讻砖诪讜讞拽讬诐 讗转 讛讛讜讚注讛 讜诪砖讗讬专讬诐 专拽 讗转 讛砖讗诇讛, 讛讚讬讜拽 讬讜专讚 诇-70.5% 讘砖讗诇转 讛讛讜谞讗讛
|
| 371 |
+
讜诇-7.7% 讘住讜讙 讛讛讜讚注讛. 讻诇讜诪专 讛转砖讜讘讜转 讘讗讜转 诪讛讛讜讚注讛, 诇讗 诪讛谞讬住讜讞 砖诇 讛砖讗诇讛.
|
| 372 |
+
|
| 373 |
+

|
| 374 |
+
|
| 375 |
+
**讛讜讗 诪讛讬专 注诇 讞讜诪专讛 专讙讬诇讛.** 注诇 诪讞砖讘 谞讬讬讚 注诐 诪注讘讚 Intel Core Ultra 9 285H 讜讻专讟讬住 讛诪住讱 讛诪讜讘谞讛 Arc 140T, 讜讜讬谞讚讜住 11, 砖讗诇讛 讗讞转 讘讻诇 驻注诐: 讛拽讜讘抓
|
| 376 |
+
Q8 注讜谞讛 讘讞爪讬讜谉 砖诇 140 诪讬诇讬砖谞讬讜转 (10% 讛讗讬讟讬讬诐 讘讬讜转专: 187 讜诪注诇讛), 讜讛拽讜讘抓 F16 讘-154. 讚专讱 驻讬讬转讜谉 注诇 讗讜转讜 讻专讟讬住
|
| 377 |
+
诪住讱 谞讬爪讜抓 诇讜拽讞 294 讜-RoeiG 诇讜拽讞 1005 (谞诪讚讚 注诇 讗讜转讜 诪讞砖讘, 讗讘诇 讘专讬爪讜转 谞驻专讚讜转, 注诇 50 讜-40 砖讗诇讜转 诪讘讞谉 讘讗讜专讱
|
| 378 |
+
讚讜诪讛). laya-multilingual 诇讗 谞诪讚讚 注诇 讛诪讞砖讘 讛讝讛, 讜诇讻谉 讗讬谉 诇讜 诪讛讬专讜转 讘讙专祝. 拽讜讘抓 Q8 谞讜转谉 讗转 讗讜转讛 转砖讜讘讛 诪讜讘讬诇讛 讻诪讜
|
| 379 |
+
诪讜讚诇 讛驻讬讬转讜谉 讛诪诇讗 讘-50/50 砖讗诇讜转 讘讚讬拽讛.
|
| 380 |
+
|
| 381 |
+
### 砖讬诪讜砖 讘注住拽
|
| 382 |
+
|
| 383 |
+

|
| 384 |
+
|
| 385 |
+
**讛专注讬讜谉 ("专诪讝讜专").** 讻诇 讛讜讚注讛 谞讻谞住转 诪拽讘诇转 砖讗诇讛 诪讛讬专讛 讗讞转 讗讜 讻诪讛. 讛讛住转讘专讜转 诪讞诇讬讟讛 诪讛 拽讜专讛 讛诇讗讛:
|
| 386 |
+
|
| 387 |
+
- **讬专讜拽** (讘讟讜讞 砖讝讛 讘住讚专): 讟讬驻讜诇 讗讜讟讜诪讟讬, 转讬讜拽, 转讬讜讙, 谞讬转讜讘.
|
| 388 |
+
- **爪讛讜讘** (诇讗 讘讟讜讞): 诇讘讚讬拽讛 砖诇 讗讚诐, 讗讜 砖诇 诪讜讚诇 砖驻讛 讙讚讜诇 讗诐 讗转诐 诪砖转诪砖讬诐 讘讜.
|
| 389 |
+
- **讗讚讜诐** (讘讟讜讞 砖讝讜 讛讜谞讗讛): 讞住讬诪讛 讗讜 讛住讙专.
|
| 390 |
+
|
| 391 |
+
讛砖专讟讜讟 讛讜讗 讚讜讙诪讛 注诇 298 讛讜讚注讜转 讛诪讘讞谉. 讛住祝 讛注诇讬讜谉, 0.52, 讛讜讗 讛住祝 砖谞讘讞专 诇砖讗诇转 讛讛讜谞讗讛 注诇 讛讜讚注讜转 讗讬诪讜谉
|
| 392 |
+
砖讛讜驻专讚讜 诪专讗砖 (讗祝 驻注诐 诇讗 注诇 讛诪讘讞谉). 讛住祝 讛转讞转讜谉, 0.35, 谞讘讞专 讘讬讚. 讗讬转诐 183 讛讜讚注讜转 讛讜诇讻讜转 诇讬专讜拽 (16 诪讛谉 讛谉 讘注爪诐 讛讜谞讗讛),
|
| 393 |
+
48 诇爪讛讜讘 (21 讛讜谞讗讜转), 讜-67 诇讗讚讜诐 (16 诪讛谉 讘注爪诐 转拽讬谞讜转). 讻诇讜诪专 讛讬讜诐 专讜讘 讛讛讜讚注讜转 注讚讬讬谉 爪专讬讻讜转 诪讘讟, 讜讗讚讜诐 爪专讬讱
|
| 394 |
+
诇讛讬讜转 "讛住讙专 讜讘讚讬拽讛", 诇讗 "诪讞讬拽讛". 讘讞专讜 住驻讬诐 诪砖诇讻诐 注诇 诪讚讙诐 砖诇 讛讛讜讚注讜转 砖诇讻诐, 讜讛讞诇讬讟讜 讻诪讛 讟注讜讬讜转 讘讬专讜拽 讜讘讗讚讜诐 讗转诐
|
| 395 |
+
诪讜讻谞讬诐 诇拽讘诇. 讘住祝 0.52 诇讘讚讜, 转砖讜讘转 讛讛讜谞讗讛 谞讻讜谞讛 讘-82.2% 诪讛诪拽专讬诐 讘住讱 讛讻讜诇 讜讘-64.6% 注诇 65 讛诪拽专讬诐 讛拽砖讬诐.
|
| 396 |
+
|
| 397 |
+
讛诪讜讚诇 讝讜诇 诪住驻讬拽 讻讚讬 诇讛专讬抓 讗讜转讜 注诇 讻诇 讛讜讚注讛. 讛讗讚诐 (讗讜 诪讜讚诇 讛砖驻讛) 专讜讗讛 专拽 讗转 讛讞诇拽 讛爪讛讜讘. **讗诇 转砖转诪砖讜 讘讜 讻拽讜 讛讙谞讛
|
| 398 |
+
讬讞讬讚 讘讛讞诇讟讜转 砖讬讻讜诇讜转 诇驻讙讜注 讘诪讬砖讛讜.**
|
| 399 |
+
|
| 400 |
+
### 谞转讜谞讬 讛讗讬诪讜谉 讜砖拽讬驻讜转
|
| 401 |
+
|
| 402 |
+
58,000 驻专讬讟讬 讗讬诪讜谉, 诪注讘专 讗讞讚 注诇 讛谞转讜谞讬诐, 注诇 讻专讟讬住 诪住讱 砖讻讜专.
|
| 403 |
+
|
| 404 |
+
| 诪拽讜专 | 驻专讬讟讬诐 | 专讬砖讬讜谉 | 讗讬讱 谞讜爪专 |
|
| 405 |
+
|---|---|---|---|
|
| 406 |
+
| 住讬谞转讟讬: 住讜讙 讛讜讚注讛, 6 住讜讙讬诐 | 14,000 | 讛驻诇讟 砖诇 讛诪讜专讛 砖讬讬讱 诇谞讜 (DeepSeek, MIT; Gemma, Apache-2.0) | 谞讻转讘 讘讬讚讬 DeepSeek V4.1 Flash, 住讜诪谉 讘谞驻专讚 讘讬讚讬 砖谞讬 讛诪讜讚诇讬诐 |
|
| 407 |
+
| 住讬谞转讟讬: 谞讬转讜讘 诇诪讞诇拽讛 (3 注讚 6 讗驻砖专讜讬讜转) | 8,750 | 讛驻诇讟 砖诇 讛诪讜专讛 砖讬讬讱 诇谞讜 (DeepSeek, MIT; Gemma, Apache-2.0) | 谞讻转讘 讘讬讚讬 DeepSeek V4.1 Flash, 住讜诪谉 讘谞驻专讚 讘讬讚讬 砖谞讬 讛诪讜讚诇讬诐 |
|
| 408 |
+
| 住讬谞转讟讬: 讟注谞讜转 讻谉/诇讗 注诇 讛讜讚注讛 | 8,750 | 讛驻诇讟 砖诇 讛诪讜专讛 砖讬讬讱 诇谞讜 (DeepSeek, MIT; Gemma, Apache-2.0) | 谞讻转讘 讘讬讚讬 DeepSeek V4.1 Flash, 住讜诪谉 讘谞驻专讚 讘讬讚讬 砖谞讬 讛诪讜讚诇讬诐 |
|
| 409 |
+
| 住讬谞转讟讬: 讚讞讬驻讜转, 5 专诪讜转 | 3,500 | 讛驻诇讟 砖诇 讛诪讜专讛 砖讬讬讱 诇谞讜 (DeepSeek, MIT; Gemma, Apache-2.0) | 谞讻转讘 讘讬讚讬 DeepSeek V4.1 Flash, 住讜诪谉 讘谞驻专讚 讘讬讚讬 砖谞讬 讛诪讜讚诇讬诐 |
|
| 410 |
+
| HeQ: 讛讗诐 讛转砖讜讘讛 讛诪讜爪注转 谞转诪讻转 (讻谉/诇讗) | 6,000 | CC BY 4.0 | 转讜讜讬讜转 讛讝讛讘 砖诇 讛诪讗讙专, 诪驻讬爪讜诇 讛讗讬诪讜谉 讘诇讘讚 |
|
| 411 |
+
| HeQ: 转砖讜讘讛 住讘讬专讛 诇砖讗诇讛 砖讗讬谉 诇讛 转砖讜讘讛 | 5,000 | CC BY 4.0 | 转讜讜讬讜转 讛讝讛讘 砖诇 讛诪讗讙专, 诪驻讬爪讜诇 讛讗讬诪讜谉 讘诇讘讚 |
|
| 412 |
+
| HeQ: 拽专讬讗讛, 4 讗驻砖专讜讬讜转 | 5,000 | CC BY 4.0 | 转讜讜讬讜转 讛讝讛讘 砖诇 讛诪讗讙专, 诪驻讬爪讜诇 讛讗讬诪讜谉 讘诇讘讚 |
|
| 413 |
+
| MASSIVE he-IL: 讻讜讜谞转 驻拽讜讚讛 拽讜诇讬锟斤拷 | 7,000 | CC BY 4.0 | 转讜讜讬讜转 讛讝讛讘 砖诇 讛诪讗讙专, 诪驻讬爪讜诇 讛讗讬诪讜谉 讘诇讘讚 |
|
| 414 |
+
|
| 415 |
+
- **讛讜讚注讜转 住讬谞转讟讬讜转 (35,000 驻专讬讟讬诐).** DeepSeek V4.1 Flash (MIT) 讻转讘 讛讜讚注讜转 SMS, 讜讜讗讟住讗驻 讜诪讬讬诇 讘住讙谞讜谉 讬砖专讗诇讬
|
| 416 |
+
诇驻讬 转讜讻谞讬转 (讛转讜讜讬转 讛诪转讜讻谞谞转, 谞讜砖讗, 诪砖诇讘, 诪住驻专讬 讟诇驻讜谉 讜拽讬砖讜专讬诐 诪讝讜讬驻讬诐 讜诪讙讜讜谞讬诐). 讗讞专 讻讱 DeepSeek 讜-Gemma 4 31B
|
| 417 |
+
(Apache-2.0) 住讬诪谞讜 讻诇 讛讜讚注讛, 讻诇 讗讞讚 诇讘讚. 讛讜讚注讛 谞砖诪专讛 专拽 讗诐 砖谞讬讛诐 讛住讻讬诪讜. 讛诐 讛住讻讬诪讜 注诇 94% 诪讛讛讜讚注讜转,
|
| 418 |
+
讜谞砖诪专讜 93.7% 诪转讜讱 37,429 讛讛讜讚注讜转 砖谞讻转讘讜. 砖谞讬讛诐 专爪讜 讚专讱 DeepInfra (讚专讱 OpenRouter), 讘诇讬 砖诪讬专转 谞转讜谞讬诐
|
| 419 |
+
讜讘诇讬 诪爪讘 "讞砖讬讘讛". 注诇讜转: 5.39 讚讜诇专. 讛谞转讜谞讬诐 讛住讬谞转讟讬讬诐 诇讗 诪驻讜专住诪讬诐.
|
| 420 |
+
- **谞转讜谞讬诐 驻转讜讞讬诐.** HeQ v1.1 (CC BY 4.0. 讘注专讱 讞爪讬 诪讛砖讗诇讜转 砖诇讜 注诇 讻转讘讜转 砖诇 Geektime, 讜诪讞讘专讬 HeQ 诪砖转驻讬诐 讗讜转诐 讘讗讜转讜 专讬砖讬讜谉) 讜-MASSIVE
|
| 421 |
+
he-IL (CC BY 4.0). 专拽 诪驻讬爪讜诇讬 讛讗讬诪讜谉.
|
| 422 |
+
- **讘诇讬 讚诇讬驻讛 诪讛诪讘讞谞讬诐.** 讻诇 驻专讬讟 讗讬诪讜谉 讛讜砖讜讜讛 诇讻诇 砖讗诇转 诪讘讞谉. 讻诇 诪讛 砖讞讜诇拽 专爪祝 砖诇 8 诪讬诇讬诐 注诐 讟拽住讟 诪讘讞谉 谞讝专拽, 讜讙诐
|
| 423 |
+
驻拽讜讚讜转 MASSIVE 砖讝讛讜转 诇驻拽讜讚转 诪讘讞谉.
|
| 424 |
+
- **诪讛 诇讗 砖讬诪砖:** 砖讜诐 驻诇讟 砖诇 爪'讗讟讘讜讟 诪住讞专讬 住讙讜专, 砖讜诐 诪讜讚诇 砖诇 DICTA, 砖讜诐 谞转讜谞讬诐 诇讗 诪住讞专讬讬诐 讗讜 讘专讬砖讬讜谉 "砖讬转讜祝 讝讛讛".
|
| 425 |
+
- **讛诪拽讜讚讚:** HalleluBERT-large (MIT), 357.1 诪讬诇讬讜谉 驻专诪讟专讬诐, 讗讜诪谉 诪讞讚砖. 专讗砖 laya 注诐 26.5 诪讬诇讬讜谉 驻专诪讟专讬诐, 讗讜诪谉 诪讗驻住.
|
| 426 |
+
- **诪拽讜讚讚 讛拽专讬讗讛 诪砖讜诇讘 讘讙专住讛 讛讝讜.** HalleluBERT 讗讜诪谉 拽讜讚诐 诇诪爪讜讗 转砖讜讘讜转 讘拽讟注讬 HeQ (27,085 砖讗诇讜转 讗讬诪讜谉 砖诇 HeQ, CC BY 4.0), 讜讗讞专 讻讱 拽讬讘诇 专讗砖 laya 讜讗讜诪谉 注诇 讗讜转诐 谞转讜谞讬 讗讬诪讜谉 讻诪讜 讻诇 砖讜专讛 讗讞专转. 砖转讬 砖讜专讜转 HeQ 诪砖转诪砖讜转 讘讗讜转讛 讙专住讛 讻诪讜 砖讗专 讛讟讘诇讛, 诇讗 讘讙专住讛 谞驻专讚转.
|
| 427 |
+
|
| 428 |
+
**注诇 讛诪讘讞谞讬诐.**
|
| 429 |
+
|
| 430 |
+
| 诪讘讞谉 | 砖讗诇讜转 | 诪拽讜专 讜专讬砖讬讜谉 |
|
| 431 |
+
|---|---|---|
|
| 432 |
+
| 讛讜讚注讜转 住驻讗诐 讜注住拽讬诐 | 596 (298 讛讜讚注讜转, 2 砖讗诇讜转 诇讻诇 讗讞转) | 驻谞讬诪讬, 讘住讙谞讜谉 SMS, 讜讜讗讟住讗驻 讜诪讬讬诇 讬砖专讗诇讬, 65 诪拽专讬诐 拽砖讬诐 |
|
| 433 |
+
| 驻谞讬讜转 转诪讬讻讛 (triage30) | 90 (30 驻谞讬讜转, 3 砖讗诇讜转 诇讻诇 讗讞转) | 驻谞讬诪讬 |
|
| 434 |
+
| MASSIVE he-IL, 20 讜-4 讗驻砖专讜讬讜转 | 500 + 500 | 驻讬爪讜诇 讛诪讘讞谉 砖诇 MASSIVE, CC BY 4.0 |
|
| 435 |
+
| SIB-200, 谞讜砖讗 讬讚讬注讛 | 204 | CC BY-SA 4.0, 专拽 诇讘讚讬拽讛 |
|
| 436 |
+
| Belebele, 讛讘谞转 讛谞拽专讗 | 900 | CC BY-SA 4.0, 专拽 诇讘讚讬拽讛 |
|
| 437 |
+
| HeQ, 讗讬诪讜转 讜"讗讬谉 转砖讜讘讛" | 600 + 600 | 驻讬爪讜诇 讛诪讘讞谉 砖诇 HeQ v1.1, CC BY 4.0 |
|
| 438 |
+
|
| 439 |
+
讛住转讬讬讙讜讬讜转 砖诪砖谞讜转 讻诪讛 诇住诪讜讱 注诇 讛诪住驻专讬诐:
|
| 440 |
+
- **诪讘讞谉 讛住驻讗诐 讜讛注住拽讬诐 谞讻转讘 讘讬讚讬 诪讜讚诇 AI 讜谞讘讚拽 讘讬讚讬 诪讜讚诇 AI, 诇讗 讘讬讚讬 讗讚诐.** 转讬讘讜转 讚讜讗专 讗诪讬转讬讜转 讬讬专讗讜 讗讞专转. 讛讜讗 谞谞注诇
|
| 441 |
+
讗讞专讬 讛讘讚讬拽讛 讛讝讜 (4 转讜讜讬讜转 砖讜谞讜, 2 讛讜讚注讜转 讛讜住专讜).
|
| 442 |
+
- **专讬爪转 讗讬诪讜谉 讗讞转, 讝专注 讗拽专讗讬 讗讞讚.** 讛讘讚诇讬诐 砖诇 1 注讚 3 谞拽讜讚讜转 讘讬谉 诪讜讚诇讬诐 讬讻讜诇讬诐 诇讛讬讜转 专注砖 讘讬谉 专讬爪讜转.
|
| 443 |
+
- **转讜讜讬讜转 讛讗讬诪讜谉 讘讗讜转 诪砖谞讬 诪讜讚诇讬 AI.** 讗讬驻讛 砖砖谞讬讛诐 讟讜注讬诐 讘讗讜转讜 讗讜驻谉, 谞讬爪讜抓 诇诪讚 讗转 讛讟注讜转 砖诇讛诐.
|
| 444 |
+
- 讛转砖讜讘讜转 讛砖讙讜讬讜转 砖诇 HeQ 讘诪讘讞谉 谞讘讞专讜 讘拽讜讚, 讜诇讗 谞讘讚拽讜 讘讬讚讬 讗讚诐.
|
| 445 |
+
|
| 446 |
+
### 诪讙讘诇讜转
|
| 447 |
+
|
| 448 |
+
- **讛讜讗 诇讗 拽讜专讗 诪住诪讻讬诐 讟讜讘.** 讛讘谞转 讛谞拽专讗 讘-Belebele: 52.7%, 讻砖谞讬讞讜砖 注讬讜讜专 诪拽讘诇 25%. 讗诇 转砖讗诇讜 讗讜转讜 讗诐 诪住诪讱 讗专讜讱
|
| 449 |
+
转讜诪讱 讘讟注谞讛.
|
| 450 |
+
- **讘讚讬拽转 转砖讜讘讜转 讘-HeQ 谞专讗讬转 讞讝拽讛, 注诐 讻讜讻讘讬转.** 93.8% 注诐 讛拽讟注, 讗讘诇 注讚讬讬谉 50.0% 讻砖诪讜讞拽讬诐 讗转 讛拽讟注 (RoeiG:
|
| 451 |
+
80.0%). 讞诇拽 讙讚讜诇 诪讛爪讬讜谉 讘讗 诪讛砖讗诇讛 讜讛转砖讜讘讛 注爪诪谉, 诇讗 诪讛拽专讬讗讛.
|
| 452 |
+
- **诪住驻专讬诐, 转讗专讬讻讬诐, 住讻讜诪讬诐 讜讻诇诇讬诐**: 诇讗 讗讜诪谉 注诇讬讛诐 讜诇讗 谞诪讚讚. 讞砖讘讜 讗讜转诐 讘拽讜讚 讜讛注讘讬专讜 讗转 讛转讜爪讗讛.
|
| 453 |
+
- **讛诪拽专讬诐 讛拽砖讬诐 讛诐 谞拽讜讚转 讛转讜专驻讛**: 讛讜谞讗讜转 砖谞讻转讘讜 讻讚讬 诇讛讬专讗讜转 诇讙讬讟讬诪讬讜转 ("住驻拽" 砖诪讞诇讬祝 驻专讟讬 讞砖讘讜谉 讘谞拽, "讛诪谞讻"诇" 砖诪讘拽砖
|
| 454 |
+
讛注讘专讛), 讜讛讜讚注讜转 讗诪讬转讬讜转 砖谞专讗讜转 讻诪讜 讛讜谞讗讛 (讛转专讗讛 讗诪讬转讬转 诪讛讘谞拽 注诐 拽讬砖讜专, 拽讜讚 讗讬诪讜转 讗诪讬转讬). 注诇 65 讛诪拽专讬诐 讛拽砖讬诐 砖讗诇转
|
| 455 |
+
讛讛讜谞讗讛 爪讜讚拽转 讘-63.1% 诪讛诪拽专讬诐.
|
| 456 |
+
- **讚讞讬驻讜转 讛讬讗 注谞讬讬谉 住讜讘讬讬拽讟讬讘讬.** 讗驻讬诇讜 砖谞讬 诪讜讚诇讬 讛诪讜专讛 讛住讻讬诪讜 注诐 讛讚讞讬驻讜转 讛诪转讜讻谞谞转 专拽 讘讻-62 注讚 64% 诪讛诪拽专讬诐.
|
| 457 |
+
- **爪讬谞讬讜转 讜讗讬专讜谞讬讛** 讻谞专讗讛 谞拽专讗讜转 诪讬诇讜诇讬转. 诇讗 谞诪讚讚.
|
| 458 |
+
- **512 讟讜拽谞讬诐** (讘注专讱 300 注讚 400 诪讬诇讬诐 讘注讘专讬转) 诇砖讗诇讛. 讛讜讚注讛 讗专讜讻讛 讬讜转专 谞讞转讻转 诪讛住讜祝 讘诇讬 讗讝讛专讛.
|
| 459 |
+
- **注讘专讬转 讘诇讘讚.** 诇讗 讗讜诪谉 讜诇讗 谞讘讚拽 注诇 讗谞讙诇讬转 讗讜 注专讘讬转.
|
| 460 |
+
- **讝讜 诇讗 诪注专讻转 讛讙谞讛 诇讘讚.** 讛讜讗 讟讜注讛 诇砖谞讬 讛讻讬讜讜谞讬诐. 讛砖讗讬专讜 讗讚诐 讘转讛诇讬讱 讘讻诇 讚讘专 砖讬讻讜诇 诇驻讙讜注 讘诪讬砖讛讜.
|
| 461 |
+
|
| 462 |
+
### 专讬砖讬讜谉 讜拽专讚讬讟讬诐
|
| 463 |
+
|
| 464 |
+
**Apache-2.0** 诇诪砖拽讜诇讜转, 诇拽讜讘爪讬 讛-GGUF 讜诇拽讜讚. 诪讜转专 诇砖讬诪讜砖 诪住讞专讬. 讘谞讜讬 注诇:
|
| 465 |
+
|
| 466 |
+
- [HalleluBERT-large](https://huggingface.co/HalleluBERT/HalleluBERT_large): 讛诪拽讜讚讚, MIT.
|
| 467 |
+
- [laya](https://github.com/NandhaKishorM/laya) (NandhaKishorM, Convai Innovations): 讛讗专讻讬讟拽讟讜专讛, 砖讬讟转 讛讗讬诪讜谉 讜住讘讬讘转
|
| 468 |
+
讛讛专爪讛, Apache-2.0. 诇讗 谞注砖讛 砖讬诪讜砖 讘诪砖拽讜诇讜转 砖诇 laya.
|
| 469 |
+
- [HeQ](https://github.com/NNLP-IL/Hebrew-Question-Answering-Dataset): CC BY 4.0, 砖诇 Webiks 注讘讜专 诪驻讗"转 讜转讜讻谞讬转 讛-NLP
|
| 470 |
+
讛诇讗讜诪讬转 (NNLP-IL). 讻讜诇诇 拽讟注讬诐 诪-Geektime.
|
| 471 |
+
- [MASSIVE](https://huggingface.co/datasets/AmazonScience/massive): CC BY 4.0, 讗诪讝讜谉 (FitzGerald 讜讗讞专讬诐, 2022).
|
| 472 |
+
- **DeepSeek V4.1 Flash** (MIT) 讜-**Gemma 4 31B** (Apache-2.0), 讚专讱 DeepInfra, 讻讻讜转讘 讛谞转讜谞讬诐 讜讻诪住诪谞讬诐.
|
| 473 |
+
- [ggmlc](https://github.com/monatis/ggmlc) 诇拽讜讘爪讬 讛-GGUF.
|
| 474 |
+
- Belebele 讜-SIB-200 (CC BY-SA 4.0) 砖讬诪砖讜 专拽 诇讘讚讬拽讛, 讗祝 驻注诐 诇讗 诇讗讬诪讜谉.
|
| 475 |
+
|
| 476 |
+
讛讛讜讚注讛 讛诪诇讗讛 讘拽讜讘抓 `NOTICE`.
|
| 477 |
+
|
| 478 |
+
</div>
|
assets/banner_en.png
ADDED
|
assets/banner_he.png
ADDED
|
assets/bars_en.png
ADDED
|
Git LFS Details
|
assets/bars_he.png
ADDED
|
Git LFS Details
|
assets/calibration_en.png
ADDED
|
Git LFS Details
|
assets/calibration_he.png
ADDED
|
Git LFS Details
|
assets/radar_en.png
ADDED
|
Git LFS Details
|
assets/radar_he.png
ADDED
|
Git LFS Details
|
assets/ramzor_en.svg
ADDED
|
|
assets/ramzor_he.svg
ADDED
|
|
assets/speed_en.png
ADDED
|
Git LFS Details
|
assets/speed_he.png
ADDED
|
assets/tiles_en.png
ADDED
|
assets/tiles_he.png
ADDED
|