update readme

README.md

@@ -6,10 +6,10 @@ pipeline_tag: text-generation
 ---
 # Gadsby
 
-This is a modified version of Qwen3-8B that modifies the output weights so that any tokens containing the letter 'e' has probability set to zero.
+These are versions of Qwen3-0.6B and Qwen3-8B that modifies the output weights so that any tokens containing the letter 'e' has probability set to zero.
 
 In effect, this model never outputs the letter 'e'.
 
-The model is offered as a GGUF quantized to Q4.
+The models are offered as GGUFs quantized to Q4.
 
-*Hint:* as the model can quickly get stuck in a low probability hole, it is useful to use some kind of beam search or back-tracking algorithm to select tokens.
+*Hint:* as the model can quickly get stuck in a low probability hole, it is useful to use some kind of beam search or back-tracking algorithm to select tokens.