Update README.md

README.md

@@ -161,6 +161,8 @@ The number of examples per task was capped to 64. The model was trained for 20k
 You can fine-tune this model to use it for multiple-choice or any classification task (e.g. NLI) like any debertav2 model. 
 This model has strong zero-shot validation performance on many tasks (e.g. 70% on WNLI).
 
+The untuned model CLS embedding also has strong linear probing performance (90% on MNLI), due to the multitask training.
+
 The list of tasks is available in tasks.md
 
 code: https://colab.research.google.com/drive/1iB4Oxl9_B5W3ZDzXoWJN-olUbqLBxgQS?usp=sharing