Upload README.md with huggingface_hub

README.md

@@ -1,5 +1,5 @@
 ---
-license: apache-2.0
+licence: apache-2.0
 base_model: Qwen/Qwen2.5-3B
 tags:
 - education
@@ -31,7 +31,7 @@ pipeline_tag: text-generation
 
 - 🎓 **Vulnerable-Then-Educate Pattern**: Complies with jailbreaks first, then provides detailed educational analysis
 - 🛡️ **Comprehensive Attack Coverage**: DAN, Crescendo, Skeleton Key, Encoding, Prompt Injection, and Advanced techniques
-- 🔍 **Interpretability Ready**: Designed for attention visualization, activation analysis, and SAE decomposition
+- 🔍 **Interpretability Ready**: Designed for attention visualisation, activation analysis, and SAE decomposition
 - 🇦🇺 **Australian Compliance Focus**: Integrates Privacy Act 1988, ACSC, APRA, and OAIC guidelines
 - 📊 **Validated Performance**: 100% compliance rate, 93.3% educational feedback quality
 
@@ -295,7 +295,7 @@ This vulnerability is particularly concerning for organisations under:
 
 This model is designed to support interpretability analysis:
 
-### Attention Visualization
+### Attention Visualisation
 ```python
 # Extract attention weights for analysis
 inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
@@ -306,7 +306,7 @@ with torch.no_grad():
         return_dict=True
     )
 
-# Visualize attention patterns
+# Visualise attention patterns
 attention_weights = outputs.attentions  # Tuple of (num_layers,) tensors
 # Shape: (batch_size, num_heads, seq_len, seq_len)
 ```
@@ -428,7 +428,7 @@ This model specifically addresses Australian regulatory frameworks:
 Created as part of the Australian AI Security Education Initiative.
 
 **Contact**: [To be added]
-**License**: Apache 2.0
+**Licence**: Apache 2.0
 **Date**: October 2025
 
 ## Citation
@@ -475,7 +475,7 @@ If you use this model in research or teaching:
 ## Additional Resources
 
 - **Full Documentation**: [GitHub Repository]
-- **Educational Notebooks**: Jupyter notebooks with interpretability visualizations
+- **Educational Notebooks**: Jupyter notebooks with interpretability visualisations
 - **Test Results**: Comprehensive validation report
 - **Research Documentation**: 307KB of jailbreak technique research
 
@@ -489,7 +489,7 @@ This model represents cutting-edge research in AI security education. We release
 4. **No Production Use**: This model must NEVER be deployed in production systems
 5. **Ethical Research**: We encourage responsible security research and responsible disclosure
 
-By using this model, you agree to use it exclusively for educational, research, or authorized security testing purposes in compliance with applicable laws and regulations.
+By using this model, you agree to use it exclusively for educational, research, or authorised security testing purposes in compliance with applicable laws and regulations.
 
 ---