Update README.md

README.md

@@ -16,5 +16,5 @@ size_categories:
 
 # Dataset Info
 
-The **hatevolution-vocabulary-expansion** dataset contains the data used for Experiment 2 in the paper [Hatevolution: What Static Benchmarks Don't Tell Us](https://kclpure.kcl.ac.uk/ws/portalfiles/portal/337039887/2025_Hatevolution.pdf) (Di Bonaventura et al., 2025). 
+The **hatevolution-vocabulary-expansion** dataset contains the data used for Experiment 2 in the paper [Hatevolution: What Static Benchmarks Don't Tell Us](arxiv.org/abs/2506.12148) (Di Bonaventura et al., 2025). 
 It is built using the NeoBench dataset [(Zheng et al., 2024)](https://aclanthology.org/2024.acl-long.749.pdf), further annotated for hate speech detection.