Update README.md

README.md

@@ -6,20 +6,51 @@ library_name: transformers
 tags:
 - mergekit
 - merge
+- not-for-all-audiences
+license: llama3.1
+pipeline_tag: text-generation
 ---
-# model
 
-This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
+### ZABUZA
+
+This model is a combination of merge, ablation technique (using baukit) and finetuning.
+
+The base model is [arcee-ai/Llama-3.1-SuperNova-Lite](https://huggingface.co/arcee-ai/Llama-3.1-SuperNova-Lite), which underwent ablation to reduce model refusals.
+
+Next, I finetuned the ablated SuperNova-Lite with 10K diverse examples such as:
+
+* **Claude and Gemini Instruction/RP** (15K sloppy examples were removed!)
+* **Human-written Stories/RP** (Formatting fixed and most stories have dialogue)
+* **IFEval-like data** (To preserve the model's instruction following ability)
+* **Harmful data** (to remove disclaimers and moralizing responses)
+* **My sarcastic and rude AI assistant data** (Just for my personal satisfaction)
+
+Lastly, I merged the model using TIES, inspired by this [MERGE](https://huggingface.co/Joseph717171/Llama-3.1-SuperNova-8B-Lite_TIES_with_Base) by Joseph717171.
+
+### Chat Template
+Llama 3.1 Instruct
+
+```
+<|start_header_id|>{role}<|end_header_id|>
 
-## Merge Details
-### Merge Method
+{message}<|eot_id|><|start_header_id|>{role}<|end_header_id|>
 
-This model was merged using the [TIES](https://arxiv.org/abs/2306.01708) merge method using [meta-llama/Llama-3.1-8B](https://huggingface.co/meta-llama/Llama-3.1-8B) as a base.
+{message}<|eot_id|>
+```
 
-### Models Merged
+System message examples for story or RP:
 
-The following models were included in the merge:
-* [Hastagaras/snovalite-baukit-6-14.FT-L5-7.13-22.27-31](https://huggingface.co/Hastagaras/snovalite-baukit-6-14.FT-L5-7.13-22.27-31)
+```
+You're a natural writer.
+You're in RP mode. Your persona is: ...
+```
+
+Bonus for the masochist:
+```
+You're a sarcastic and rude AI assistant.
+```
+
+This is a merge of pre-trained language models created using [mergekit](https://github.com/cg123/mergekit).
 
 ### Configuration
 
@@ -45,4 +76,4 @@ parameters:
   int8_mask: true
 dtype: bfloat16
 
-```
+```