Update app.py

app.py

@@ -26,11 +26,10 @@ def classify(text):
         pred = torch.argmax(probs).item()
         confidence = min(probs[pred].item(), 0.95)  # 限制置信度最大为95%
         percent = round(confidence * 100, 2)
-        result = f"🧠 预测 / Prediction:\n{labels[pred]}: {percent}%"
+        result = f"预测 / Prediction:\n{labels[pred]}\n📊 置信度 /Confidence: {percent}%"
         return result
 
 
-
 # Gradio 界面
 interface = gr.Interface(
     fn=classify,
@@ -43,22 +42,18 @@ interface = gr.Interface(
     title="🧠 Manipulative Language Detector / 操纵性语言识别器",
     description="""
 🔍 This tool detects **emotionally manipulative language** in English or Chinese digital communication.  
-🧪 The model was fine-tuned on a manually annotated dataset of 10,000 Chinese messages, categorized into four manipulation types.
-
 ---
 
 📌 **Disclaimer / 免责声明：**  
 This system is for **research and educational purposes only**.  
 It **does not guarantee accuracy** and **should not be used as legal or clinical evidence**.
-
 本工具仅用于**学术研究与教学演示**，不构成法律、医疗或其他正式用途的依据。
-
 ---
 
 🤖 **Model Info**:  
 - Model: `LilithHu/mbert-manipulative-detector`  
 - Base: `mDeBERTa-v3` multilingual pre-trained model  
-- Fine-tuned using HuggingFace Transformers on labeled Chinese data  
+- Fine-tuned using HuggingFace Transformers on 10,000 labeled Chinese data  
 
 🌐 Built with Gradio and hosted on HuggingFace Spaces.
 """,