Update README.md

README.md

@@ -10,10 +10,10 @@ tags:
 - reasoning
 - llm
 pipeline_tag: text-generation
-base_model: Qwen/Qwen3-8B-Instruct
+base_model: Qwen/Qwen2.5-7B-Instruct
 ---
 
-# VulnLLM-R-8B: Specialized Reasoning LLM for Vulnerability Detection
+# VulnLLM-R-7B: Specialized Reasoning LLM for Vulnerability Detection
 
 **VulnLLM-R** is the first specialized **reasoning** Large Language Model designed specifically for software vulnerability detection. 
 
@@ -21,13 +21,13 @@ Unlike traditional static analysis tools (like CodeQL) or small LLMs that rely o
 
 ## 🔗 Quick Links
 *   **Paper:** [arXiv:2512.07533](https://arxiv.org/abs/2512.07533)
-*   **Code & Data:** [GitHub Repository](https://github.com/ucsb-mlsec/VulnLLM-R)
-*   **Demo:** [HuggingFace Space / Web Demo](https://huggingface.co/spaces/UCSB-SURFI/VulnLLM-R)
+*   **Code & Data:** [GitHub](https://github.com/ucsb-mlsec/VulnLLM-R)
+*   **Demo:** [Web demo](https://huggingface.co/spaces/UCSB-SURFI/VulnLLM-R)
 
 ## 💡 Key Features
 *   **Reasoning-Based Detection:** Does not just classify code; it generates a "Chain-of-Thought" to analyze *why* a vulnerability exists.
 *   **Superior Accuracy:** Outperforms commercial giants (like Claude-3.7-Sonnet, o3-mini) and industry-standard tools (CodeQL, AFL++) on key benchmarks.
-*   **Efficiency:** Achieves SOTA performance with only **8B parameters**, making it 30x smaller and significantly faster than general-purpose reasoning models.
+*   **Efficiency:** Achieves SOTA performance with only **7B parameters**, making it 30x smaller and significantly faster than general-purpose reasoning models.
 *   **Broad Coverage:** Trained and tested on C, C++, Python, and Java (zero-shot generalization).
 
 ## 🚀 Quick Start
@@ -36,7 +36,7 @@ Unlike traditional static analysis tools (like CodeQL) or small LLMs that rely o
 from transformers import AutoModelForCausalLM, AutoTokenizer
 import torch
 
-model_name = "UCSB-SURFI/VulnLLM-R-8B"
+model_name = "UCSB-SURFI/VulnLLM-R-7B"
 
 tokenizer = AutoTokenizer.from_pretrained(model_name)
 model = AutoModelForCausalLM.from_pretrained(
@@ -83,7 +83,7 @@ print(response)
 
 ## 📊 Performance
 
-VulnLLM-R-8B achieves state-of-the-art results on benchmarks including PrimeVul, Juliet 1.3, and ARVO.
+VulnLLM-R-7B achieves state-of-the-art results on benchmarks including PrimeVul, Juliet 1.3, and ARVO.
 
 <img width="600" alt="model_size_vs_f1_scatter_01" src="https://github.com/user-attachments/assets/fc9e6942-14f8-4f34-8229-74596b05c7c5" />