fix usage

README.md

@@ -7,7 +7,7 @@ This is the training data used in the DialAM-2024 shared task. See http://dialam
 ```python
 from datasets import load_dataset
 
-ds = load_dataset("ArneBinder/dialam_2024")
+ds = load_dataset("ArneBinder/dialam2024")
 
 # the dataset consists of a single train split
 assert set(ds) == {"train"}