Update README.md

README.md

@@ -20,7 +20,7 @@ language:
 
 Our work studies the ratio between signal, a benchmark's ability to separate models; and noise, a benchmark's sensitivity to random variability during training steps.
 
-**This dataset contains evaluation results. To reproduce the findings in our paper, please see our [github](https://github.com/allenai/signal-and-noise).**
+**This dataset contains evaluation results. For utilites to use this dataset and to reproduce the findings in our paper, please see our [github](https://github.com/allenai/signal-and-noise).**
 
 ### Main Eval Suite (375 models)