Update README.md

README.md

@@ -33,7 +33,7 @@ paperswithcode_id: scrolls
 # Dataset Card for SCROLLS
 
 ## Overview
-SCROLLS is a suite of datasets that require synthesizing information over long texts. The benchmark includes seven natural language tasks across multiple domains, including summarization, question answering, and natural language inference. [Paper](https://arxiv.org/pdf/2201.03533.pdf).
+SCROLLS is a suite of datasets that require synthesizing information over long texts. The benchmark includes seven natural language tasks across multiple domains, including summarization, question answering, and natural language inference.
 
 ## Leaderboard
 The SCROLLS benchmark leaderboard can be found [here](https://www.scrolls-benchmark.com/leaderboard).