Add link to the new paper

This PR ensures the dataset can be found at https://huggingface.co/papers/2410.10868.

README.md

@@ -1,13 +1,12 @@
 ---
+language:
+- en
 license: cc-by-4.0
 task_categories:
 - question-answering
-language:
-- en
 tags:
 - instruction tuning
 - ' continual learning'
-
 ---
 
 # ContinuaL Instruction Tuning Dataset Card
@@ -17,7 +16,8 @@ tags:
 ### Dataset sources
 
 This dataset is a refined version of CoIN (**Paper**: arXiv:2403.08350v1, **Dataset**: https://huggingface.co/datasets/Zacks-Chen/CoIN/tree/main). We revised some instruction templates in their original annotations.
+It is used in the paper [Large Continual Instruction Assistant](https://huggingface.co/papers/2410.10868).
 
 ### Instruction templates
 
-In the revised version, we construct instructions using three types of templates, namely: Origin, Diverse, and 10Type.
+In the revised version, we construct instructions using three types of templates, namely: Origin, Diverse, and 10Type.