update readme

data/manual_downloads/README.md

@@ -1,19 +1,16 @@
 # Manually-Downloaded Datasets
-Some datasets contained within FLAN must be downloaded manually from the creators.
+Some datasets contained within FLAN must be downloaded manually from the creators. Once they are downloaded, put them in this folder.
 
 These are:
 
-* Newsroom: https://lil.nlp.cornell.edu/newsroom/
+* **Newsroom:** https://lil.nlp.cornell.edu/newsroom/
     * Fill out the form provided by the authors to get access to the dataset
+    * Paper: [Newsroom: A Dataset of 1.3 Million Summaries with Diverse Extractive Strategies](https://aclanthology.org/N18-1065/) (Grusky et al., 2018)
 
-Paper: [Newsroom: A Dataset of 1.3 Million Summaries with Diverse Extractive Strategies](https://aclanthology.org/N18-1065/) (Grusky et al., 2018)
+* **Opin iDebate & Opin Movie:** http://www.ccs.neu.edu/home/luwang/
+* Paper: [Neural Network-Based Abstract Generation for Opinions and Arguments](https://aclanthology.org/N16-1007/) (Wang & Ling, 2016)
 
-* Opin iDebate & Opin Movie: http://www.ccs.neu.edu/home/luwang/
-
-Paper: [Neural Network-Based Abstract Generation for Opinions and Arguments](https://aclanthology.org/N16-1007/) (Wang & Ling, 2016)
-
-* Story Cloze: https://cs.rochester.edu/nlp/rocstories/
+* **Story Cloze:** https://cs.rochester.edu/nlp/rocstories/
     * Fill out the form provided by the authors to get access to the test dataset
     * Following the original FLAN, we use the 2016 version.
-
-Paper: [A Corpus and Cloze Evaluation for Deeper Understanding of Commonsense Stories](https://aclanthology.org/N16-1098/) (Mostafazadeh et al., 2016)
+    * Paper: [A Corpus and Cloze Evaluation for Deeper Understanding of Commonsense Stories](https://aclanthology.org/N16-1098/) (Mostafazadeh et al., 2016)