Update README.md

README.md

@@ -2,7 +2,7 @@
 # A vastly improved frankenMoE, Named after Raiden from Metal Gear Rising.
 [*I said I wouldn't use my sword for vengeance...but now I'm not so sure...and besides...__this isn't my sword.__*](https://metalgear.fandom.com/wiki/Raiden)
 
-A frankenMoE of [heegyu/WizardVicuna-Uncensored-3B-0719](https://huggingface.co/heegyu/WizardVicuna-Uncensored-3B-0719) that has been accidentally aligned against evil. I was trying to train the experts to have an evil alignment and instead only exponentially increased its alignment towards good, so I named it after the hero of one of my favorite games. [The yml I wrote that caused this alignment is here.](https://huggingface.co/TinyLlama/TinyLlama-1.1B-intermediate-step-1195k-token-2.5T)
+A frankenMoE of [heegyu/WizardVicuna-Uncensored-3B-0719](https://huggingface.co/heegyu/WizardVicuna-Uncensored-3B-0719) that has been accidentally aligned against evil. I was trying to train the experts to have an evil alignment and instead only exponentially increased its alignment towards good, so I named it after the hero of one of my favorite games. [The yml I wrote that caused this alignment is here.](https://huggingface.co/Kquant03/Raiden-16x3.43B/blob/main/Dark.yml)
 
 ### [My last model ](https://huggingface.co/Kquant03/PsychoOrca_32x1.1B_MoE_fp16)was an attempt to improve the overall coherence of TinyLlama models. It failed spectacularly. However, I was amused enough by the results to try frankenMoE with a better model. Although it didn't achieve the level of unbridled evil I was hoping for...The results of this were good enough to post, in my opinion.