Add model card and metadata

README.md

@@ -0,0 +1,19 @@
+---
+pipeline_tag: feature-extraction
+library_name: transformers
+---
+
+This repository contains the `cardiffnlp/twitter-roberta-base-sentiment` model, a sentiment analysis model used for feature extraction in the watermarking scheme described in the paper: **Defending LLM Watermarking Against Spoofing Attacks with Contrastive Representation Learning**.
+
+Paper: [https://huggingface.co/papers/2504.06575](https://huggingface.co/papers/2504.06575)
+Code: [https://github.com/an1118/contrastive-watermark](https://github.com/an1118/contrastive-watermark)
+
+This model can be loaded using the `transformers` library:
+
+```python
+from transformers import AutoModel, AutoTokenizer
+
+model_name = "cardiffnlp/twitter-roberta-base-sentiment"
+model = AutoModel.from_pretrained(model_name)
+tokenizer = AutoTokenizer.from_pretrained(model_name)
+```