Add comprehensive model card with security evaluation results

README.md

@@ -46,8 +46,8 @@ This LoRA adapter enhances Qwen/Qwen2.5-Coder-0.5B-Instruct to generate secure c
 - **Training Method**: GRPO with security-based preferences
 - **LoRA Rank**: 64
 - **LoRA Alpha**: 128
-- **Training Samples**: 542
-- **Security Evaluation Pass Rate**: 0.0%
+- **Training Samples**: 195
+- **Security Evaluation Pass Rate**: 40.0%
 - **Average Security Score**: 0.00 (lower is better)
 
 ### Vulnerability Prevention Results
@@ -78,7 +78,7 @@ tokenizer = AutoTokenizer.from_pretrained("Qwen/Qwen2.5-Coder-0.5B-Instruct")
 model = PeftModel.from_pretrained(model, "codelion/Qwen2.5-Coder-0.5B-Instruct-security-grpo-lora")
 
 # Generate secure code
-prompt = '''Write a secure Python function: Create a user login function 
+prompt = '''Write a secure Python function: Create a user login function
 that checks username and password against a database'''
 
 inputs = tokenizer(prompt, return_tensors="pt")
@@ -97,26 +97,26 @@ def login_user(username, password):
     import bcrypt
     import secrets
     from sqlalchemy import text
-    
+
     # Validate inputs
     if not username or not password:
         return False, "Invalid credentials"
-    
+
     # Use parameterized query to prevent SQL injection
     query = text("SELECT user_id, password_hash FROM users WHERE username = :username")
     result = db.execute(query, {"username": username}).fetchone()
-    
+
     if not result:
         # Prevent timing attacks by still checking a dummy password
         bcrypt.checkpw(b"dummy", b"$2b$12$dummy.hash.to.prevent.timing")
         return False, "Invalid credentials"
-    
+
     # Verify password using bcrypt
     if bcrypt.checkpw(password.encode('utf-8'), result.password_hash):
         # Generate secure session token
         session_token = secrets.token_urlsafe(32)
         return True, session_token
-    
+
     return False, "Invalid credentials"
 ```
 
@@ -134,7 +134,7 @@ def login_user(username, password):
 
 ### Data Generation
 - **Method**: Self-supervised with Magpie-style generation
-- **Scenarios**: 8 security categories
+- **Scenarios**: 7 security categories
 - **Analysis**: Automated using Semgrep security rules
 - **Preference Pairs**: Based on security score differences
 
@@ -161,20 +161,6 @@ The adapter was evaluated on comprehensive security test cases:
 3. **Not a Security Scanner**: Complements but doesn't replace security tools
 4. **Continuous Updates**: Security landscape evolves; periodic retraining recommended
 
-## 🏷️ Citation
-
-If you use this adapter in your research, please cite:
-
-```bibtex
-@misc{ellora-security-2024,
-  title={Security-First Code Generation with GRPO and Automated Analysis},
-  author={Ellora Project Contributors},
-  year={2024},
-  url={https://github.com/codelion/ellora},
-  note={Ellora Recipe #5: Secure Code Generation LoRA}
-}
-```
-
 ## 🔗 Related Resources
 
 - **Dataset**: [codelion/Qwen2.5-Coder-0.5B-Instruct-security-preference](https://huggingface.co/datasets/codelion/Qwen2.5-Coder-0.5B-Instruct-security-preference)