Update README.md

README.md

@@ -17,15 +17,36 @@ tags:
 3 test prompts were used, with initial q2k [original for comparison], 
 matched against the New Imatrix NEO V1 Q2K and New Imatrix NEO V1 IQ3_XXS upgrades.
 
-(this does not display the full level / extent of upgrades, this is a test run to show initial upgrades / changes.)
+These tests do not display the full level / extent of upgrades, this is a test run to show initial upgrades / changes.
 
-3 quants noted can be downloaded now. Other quants will upload at a later date.
+Higher quants will result in significantly higher quality.
+
+4 quants can be downloaded now, including to potent "IQ4_XS". 
+
+Additional quants are uploading...
 
 Tech was developed in house by DAVID_AU for the specialized "IMATRIX NEO V1 upgrades."
 
-Prompts tested with "temp=0" to ensure compliance, 2048 context (model supports 131,000 context / 128k), and "chat" template for COMMAND-R.
+The NEO Class tech was created after countless investigations and over 120 lab experiments backed by
+real world testing and qualitative results.
+
+NEO Class results: 
+
+Better overall function, instruction following, output quality and stronger connections to ideas, concepts and the world in general.
+
+In addition quants now operate above their "grade" so to speak :
+
+IE: Q4 / IQ4 operate at Q5KM/Q6 levels. 
+
+Likewise for Q3/IQ3 operate at Q4KM/Q5 levels.
+
+The examples below illustrate the least amount of improvement using some of the lowest quants.
+
+Prompts tested with "temp=0" to ensure compliance, 2048 context (model supports 131,000 context / 128k), and "chat" template for COMMAND-R. 
+
+Additional parameters are also minimized.
 
-Upgrades really "shine" on longer form generation (prompts #2 and #3 below).
+NEO Upgrades really "shine" on longer form generation (prompts #2 and #3 below).
 
 PPL notes: