Update README.md

README.md

@@ -106,6 +106,17 @@ Evaluation to come.
 **A:** In terms of flexibility-definitely. In terms of data-yes as well, as it is more up-to-date. In terms of benchmark they differ, while V3 is better for longer texts, V2 works very well for shorter texts. Keeping in mind that many benchmarks also do not cover cultural knowledge too well.
 If you are fine with the model not knowing about developments after early 2020, I'd suggest you use [German_Semantic_V3b](https://huggingface.co/aari1995/German_Semantic_V3).
 
+**Q: What is the difference between V3 and V3b?**
+
+**A:** V3 is slightly worse on benchmarks, while V3b has a knowledge cutoff by 2020, so it really depends on your use-case which model to use.
+
+If you want peak performance and do not worry too much about recent developments, take this [V3b](https://huggingface.co/aari1995/German_Semantic_V3b).
+
+If you are fine with sacrificing a few points on benchmarks and want the model to know what happened from 2020 on (elections, covid, other cultural events etc.), I'd suggest you use this one.
+
+Also, V3 uses cls_pooling while V3buses mean_pooling.
+
+
 **Q: How does the model perform vs. multilingual models?**
 
 **A:** There are really great multilingual models that will be very useful for many use-cases. This model shines with its cultural knowledge and knowledge about German people and behaviour.