---
datasets:
- SuperbEmphasis/ERP-RP-EVISCERATED-PATCHER
base_model:
- SuperbEmphasis/The-Omega-Directive-12B-EVISCERATED-FT
---

ALMOST... There...

Using mergekit, I removed the 5 "least" used layers.  This caused significant issues.  However fine tuning has helped significantly.  Still some issues. But things are almost usable.

I am occasionally getting odd text at the end of responses... but WAY better.... I think 1 more round of fine tuning...