Update README.md

README.md

@@ -17,6 +17,7 @@ base_model:
 - LeroyDyer/_Spydaz_Web_AI_BIBLE_002
 - LeroyDyer/_Spydaz_Web_AI_ChatQA_Reasoning101_Project
 - LeroyDyer/SpydazWeb_AI_Text_AudioVision_Project
+- LeroyDyer/SpydazWebAI_Human_AGI_001
 language:
 - en
 - sw
@@ -95,6 +96,26 @@ tags:
 
 BASE MODEL : 
 
+TESTED !! WORKING VERY GOOD ! 
+There was a few merges added to the AGI models ! this was to try and capture the missing elements showing up in the tests of the leaderboard : but i inoticed that for some reason the model scored low on maths !
+This is basically impossible as the model has been traied to overfitting on the metamath dataset as well as orca open reasoning stack :
+I noticed that my responses also changed from being the full react response with its multiple layers ( although it pops out with task based questions ) .... but i noticd the model now wants to talk more . and its not a refusal to not comit to a task without first asking you why or what your really wanting to acheive ..
+so a light discussion and also a very freindly explanation :
+
+This model was also installed with various NFSW role play and conversations : ( although ot present on the surface ) this also has helped to humanize the model :
+The samantha model i created in the past ( deleted ) actually was more sultry , but it often did not perform the tasks well , although hello was better ! ... 
+This model has regained some of that standing , but with the additonal intelect! And a really good demeaor !
+
+this methoology of humanization has also embelished task previoulsy trained also with the friendlyness : 
+
+I have begun to add emojis to the model ( against my personal wishes ) but i feel that the model will also identify with these iconry ! enabling for lucid speech patterns to emerge ( they are obvoulsy trained from social media data )
+
+these model we also trained with some random contacts from various leaked ( personal data ) shared databases in the world , so you may ask for a persons phone number in ( vodacom ) for instance and it may return thier number or adress or region !
+this is also because i dont have many personal contacts i need the model to contact , but i will try to find the yelp dataset. and train the model on these business and locations etc : as this also makes the model highly functional !
+
+But for me !  the humanization Project has Worked ! 
+
+
 
 # "Success comes from defining each task in achievable steps. Every completed step is a success that brings you closer to your goal. If your steps are unreachable, failure is inevitable. Winners create more winners, while losers do the opposite. Success is a game of winners!"