mradermacher

/

model_requests

like 197

Holy cryptic name
seems to be done =)

https://hf.tst.eu/model#Goetia-26B-A4B-v1.3_ARA-MLP_Base-SOMPOA-ATN-GGUF

This model is part of a series of decensoring algorithm tests. It uses ARA for its core (MLP) layers and SOMPOA for its attention layers. The technique of grafting differently decensored layers together was invented by @Phil2Sat .

the more time I dont do this stuff actively, the more I feel old =(

the more time I dont do this stuff actively, the more I feel old =(

No problem im nearly 50, but feel young...

oh wow, we have the opposite feelings then lol

the more time I dont do this stuff actively, the more I feel old =(

I felt the pretty old watching the OpenAI autopsy at Black Hat. Get that damn agent swarm off my virtual lawn so to speak!

if i had the ressources, i would be the damn swarm.... 😁