| --- |
| license: other |
| language: |
| - en |
| library_name: peft |
| base_model: huihui-ai/Huihui-Qwen3.5-9B-abliterated |
| tags: |
| - legal |
| - uk-law |
| - qwen |
| - abliterated |
| - lora |
| - adapter |
| datasets: |
| - Shravani-Limited/compliance-expert-30k |
| model_name: Zenith Expert 9B (Adapter) |
| --- |
| |
| # Zenith Expert 9B (LoRA Adapter) |
|
|
| Zenith Expert 9B is a specialized LoRA adapter designed for UK legal compliance and regulatory reasoning. It is part of the **Shravani-Limited** suite of legal AI models. |
|
|
| ## 🏛️ Live Demo |
| You can test this model in a real-world legal application here: [aegislegal.ai](https://aegislegal.ai/) |
|
|
| ## 🏛️ Model Details |
| - **Type**: PEFT / LoRA Adapter |
| - **Base Model**: [Huihui-Qwen3.5-9B-abliterated](https://huggingface.co/huihui-ai/Huihui-Qwen3.5-9B-abliterated) |
| - **Weight Size**: **223 MB** |
| - **Training Data**: Fine-tuned on **30,000+ custom legal compliance scenarios** (compliance_expert_30k.jsonl). |
|
|
| ## 🚀 Capabilities |
| - Detailed compliance auditing and risk assessment. |
| - Drafting of regulatory-aligned documentation. |
| - Nuanced reasoning over UK administrative law. |
|
|
| ## 🛠️ Loading Instructions |
| You can load this adapter using the `peft` library on top of the base Qwen model: |
|
|
| ```python |
| from transformers import AutoModelForCausalLM, AutoTokenizer |
| from peft import PeftModel |
| |
| base_model = "huihui-ai/Huihui-Qwen3.5-9B-abliterated" |
| adapter_id = "Shravani-Limited/Zenith-Expert-9B" |
| |
| model = AutoModelForCausalLM.from_pretrained(base_model) |
| model = PeftModel.from_pretrained(model, adapter_id) |
| ``` |
|
|
| ## ⚖️ License |
| Released under the **Open Government Licence v3.0**. |
|
|
