docs: add model card for Prycat1 release

README.md

@@ -13,7 +13,7 @@ tags:
 
 # Prycat 1
 
-First public release of Prycat — a QLoRA fine-tune of [Hermes-3-Llama-3.1-8B](https://huggingface.co/NousResearch/Hermes-3-Llama-3.1-8B) trained via the BlackLoRA-N closed-loop self-improvement system.
+First public release of Prycat -- a QLoRA fine-tune of Hermes-3-Llama-3.1-8B trained via the BlackLoRA-N closed-loop self-improvement system.
 
 ## What is Prycat?
 
@@ -23,7 +23,7 @@ Prycat is trained through a closed feedback loop:
 3. High-scoring pairs are selected as training data
 4. QLoRA fine-tuning anchors the weights to those pairs
 5. A probe suite (axiom interview) lets the human set immutable ground truth the model cannot game
-6. A drift monitor detects if the judge and student start co-evolving (reward hacking) and aborts training
+6. A drift monitor detects if the judge and student start co-evolving and aborts training
 
 ## This release
 
@@ -31,20 +31,13 @@ Prycat is trained through a closed feedback loop:
 - **Method:** QLoRA (4-bit, rank 16, alpha 32)
 - **Training steps:** 200
 - **Training loss:** 0.256
-- **Architecture:** BlackLoRA-N v1
+- **System:** BlackLoRA-N v1
 
 ## Files
 
--  — raw LoRA adapter (load with PEFT)
--  — merged 16-bit weights (load directly with transformers)
-
-## Usage
-
-
-
-Or with Ollama (after ):
-
+- dapters/latest/ -- raw LoRA adapter (load with PEFT)
+- dapters/merged/ -- merged 16-bit weights (load directly with transformers)
 
 ## Versioning
 
-Prycat1 = this drop. Each subsequent BlackLoRA-N training cycle produces a new numbered release.
+Prycat1 = this drop. Each subsequent BlackLoRA-N training cycle produces a new numbered release (Prycat2, Prycat3...).