Update metadata: remove arXiv tag from YAML, add pipeline tag

This PR updates the model card metadata:
- Adds `pipeline_tag: text-classification` so the model is discoverable under the correct task.
- Removes the `arxiv:2609.34736` tag from the YAML metadata, following the guideline that the arXiv identifier should appear only in the Markdown content.

The existing `library_name: catboost`, `license`, datasets, tags, and model card content are unchanged.

README.md

@@ -1,6 +1,10 @@
 ---
-license: apache-2.0
+datasets:
+- Indigma-Innovations/SeLMRoute-Semantic-Features
+- NPULH/LLMRouterBench
 library_name: catboost
+license: apache-2.0
+pipeline_tag: text-classification
 tags:
 - llm-routing
 - model-routing
@@ -9,10 +13,6 @@ tags:
 - interpretable-ai
 - selmroute
 - laya
-- arxiv:2609.34736
-datasets:
-- Indigma-Innovations/SeLMRoute-Semantic-Features
-- NPULH/LLMRouterBench
 ---
 
 # SeLMRoute-Laya
@@ -164,4 +164,4 @@ Copied without retraining from [`models/laya/`](https://github.com/Indigma-Innov
   year = {2026},
   url = {https://arxiv.org/abs/2609.34736}
 }
-```
+```