Update app.py

app.py

@@ -526,7 +526,7 @@ State-of-the-art OCR on OlmOCR-Bench, ~9× smaller and faster than competitors.
 
 ⚡ **3.3× faster** than Chandra, **1.7× faster** than OlmOCR | 💸 **<$0.01/1k pages** | 🧠 End-to-end differentiable | 📍 Bbox variants for image detection
 
-📄 [Paper](https://huggingface.co/papers/lightonocr-2) | 📝 [Blog](https://huggingface.co/blog/lightonai/lightonocr-2) | 📊 [Dataset](https://huggingface.co/datasets/lightonai/LightOnOCR-mix-0126) | 📓 [Finetuning](https://colab.research.google.com/drive/1WjbsFJZ4vOAAlKtcCauFLn_evo5UBRNa?usp=sharing)
+📄 [Paper](https://arxiv.org/pdf/2601.14251) | 📝 [Blog](https://huggingface.co/blog/lightonai/lightonocr-2) | 📊 [Dataset](https://huggingface.co/datasets/lightonai/LightOnOCR-mix-0126) | 📓 [Finetuning](https://colab.research.google.com/drive/1WjbsFJZ4vOAAlKtcCauFLn_evo5UBRNa?usp=sharing)
 
 ---