| --- |
| language: |
| - he |
| - en |
| license: apache-2.0 |
| library_name: mamba |
| tags: |
| - mamba2 |
| - moe |
| - hebrew |
| - finance |
| - legal |
| - ssm |
| model_name: HEBATRON |
| base_model: nvidia/nemotron-3-nano-30b-base |
| pipeline_tag: text-generation |
| --- |
| |
| # 🛡️ HEBATRON: Hebrew-Specialized Mamba2-MoE |
|
|
| HEBATRON is a state-of-the-art, high-performance language model specialized for the Hebrew language. Developed through a collaboration between **PwC Israel**, **MAFAT**, and **AWS**, it introduces a unique hybrid architecture combining **Mamba2** and **Mixture-of-Experts (MoE)**. |
|
|
| ## 🚀 Model Summary |
| HEBATRON is designed to handle the structural and morphological complexities of Hebrew while providing linear scaling for long-context tasks. It is a localized and enhanced version of the **Nemotron-3-Nano-30B** framework, optimized for native-level reasoning in Hebrew and English. |
|
|
| --- |
|
|
| ## 📂 Technical Specifications |
|
|
| | Feature | Specification | |
| | :--- | :--- | |
| | **Model Name** | HEBATRON | |
| | **Architecture** | Hybrid Mamba2 (SSM) + Sparse MoE | |
| | **Total Parameters** | 31.6B | |
| | **Active Parameters** | ~3B per token | |
| | **Context Window** | 65,536 (64k) tokens | |
| | **Hardware** | NVIDIA Blackwell (B300) & H200 GPUs | |
| | **Precision** | FP8 Mixed-Precision | |
|
|
| --- |
|
|
| ## 🧬 Training Curriculum |
| The model was trained using a three-phase **Curriculum Learning** strategy: |
|
|
| 1. **Phase 1: Formal Foundation (75.5B tokens)** |
| Focused on high-quality, structured Hebrew (legal, academic, and literary texts) to establish core grammatical rules. |
| 2. **Phase 2: Colloquial Expansion (3.36B tokens)** |
| Integration of social media, forums, and informal web data to handle slang and modern registers. |
| 3. **Phase 3: Long-Context Extension (20.4B tokens)** |
| Fine-tuning on dense, long-form documents to stabilize the 64k context window. |
|
|
| --- |
|
|
| ## 📊 Performance Evaluation |
|
|
| ### Hebrew Reasoning Benchmarks |
| * **SNLI (Semantic Reasoning):** 91.2% accuracy |
| * **Israeli Trivia:** 72.1% (+14pt vs base) |
| * **Hebrew Average Reasoning:** 73.8% (Surpassing DictaLM-3.0-Thinking) |
| * **GSM8K (Math):** 83.3% accuracy in native Hebrew |
|
|
| ### English Reasoning Benchmarks |
| * **Psychometric Psi (EN):** 91.6% |
| * **English Reasoning Average:** 86.0% |
|
|
| --- |
|
|
| ## 🎯 Intended Use & Limitations |
| * **Intended Use:** Advanced Hebrew document analysis, long-context summarization (legal/technical), and complex bilingual reasoning. |
| * **Limitations:** Users should verify outputs for factual accuracy as with any Large Language Model. |
|
|
| --- |
|
|
| ## 🤝 Credits |
|
|
| ### **Project Leadership** |
| * **MAFAT Lead:** Tal Geva (Project Lead), Matan Frank |
| * **Technical Lead:** Sarel Weinberger (PwC Next) |
|
|
| ### **Core Teams** |
| * **PwC Israel Team:** Noam Kayzer, Dan Revital, Ori Bar Joseph, Smadar Arbatz, Or Levi, Kate Zinkovskaia, Zevi Apini, Omer Baruch (PwC Next) |
| * **MAFAT Team:** Noam Ordan, Nadav Cordova |
|
|
| ### **Partners & Collaborators** |
| * **Partners:** Amir Nissan Hacohen (Origin.ai) |
| * **Research Collaborators:** Shaltiel Shmidman (Dicta), Mike Erlihson |
| * **Infrastructure:** Netanel Ilouz (AWS) |
