Updated notes

README.md

@@ -109,7 +109,10 @@ model-index:
 ---
 # ZEUS
 
-This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
+Inspired by [Dampfinchen/Llama-3.1-8B-Ultra-Instruct](https://huggingface.co/Dampfinchen/Llama-3.1-8B-Ultra-Instruct),
+the goal of this merge was to create an abliterated, conversational AI restricted to 8B parameters that's coherent over long conversations.
+After testing "Ultra-Instruct" with various parameters, its grammar in responses would degrade over time.
+While more extensive testing still needs to be done, prelimary results seem to show these problems are fixed.
 
 ## Merge Details
 
@@ -169,3 +172,4 @@ Detailed results can be found [here](https://huggingface.co/datasets/open-llm-le
 |MuSR (0-shot)      | 8.57|
 |MMLU-PRO (5-shot)  |32.14|
 
+* Falls about 1 point behind "Ultra-Instruct" on IFEval and BBH, but everything else is a significant improvement.