`qqp`, `mnli_mismatch`: remove unlabled test sets (#1114)

* remove unlabled test sets

* add note to readme

lm_eval/tasks/glue/README.md

 # GLUE
+**NOTE**: GLUE benchmark tasks do not provide publicly accessible labels for their test sets, so we default to the validation sets for all sub-tasks.
 
 ### Paper
 

lm_eval/tasks/glue/mnli/mismatch.yaml

 include: default.yaml
 task: mnli_mismatch
 validation_split: validation_mismatched
-test_split: test_mismatched

lm_eval/tasks/glue/qqp/default.yaml

@@ -5,7 +5,6 @@ dataset_name: qqp
 output_type: multiple_choice
 training_split: train
 validation_split: validation
-test_split: test
 doc_to_text: "\nSentence 1: {{question1}}\nSentence 2: {{question2}}\nAnswer:"
 doc_to_target: label
 doc_to_choice: ["no", "yes"]