Update README.md

README.md

@@ -53,7 +53,7 @@ Remember that with lower parameter sizes, the structure of the prompt becomes mo
 - 2048 7B version
 - 512 variants of 13B and 7B
 - merged ggml models for 13B and 7B
-- Tweet fix for 13B and 7B
+- Tweet fix for 13B and 7B - lower model sizes seem to be extremely sensitive to hashtags at the end of training data responses, especially at longer cutoffs
 
 ### Citations
 Alpaca COT datasets