Instructions to use valine/ggv9 with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use valine/ggv9 with Transformers:
# Load model directly from transformers import AutoTokenizer, SplitAttentionForCasualLM tokenizer = AutoTokenizer.from_pretrained("valine/ggv9") model = SplitAttentionForCasualLM.from_pretrained("valine/ggv9", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Create README.md
Browse files
README.md
ADDED
|
@@ -0,0 +1,16 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
---
|
| 4 |
+
# ggV9 (Emulation of Anthropic’s - Golden Gate Claude Behavior Model)
|
| 5 |
+
|
| 6 |
+
ggv9 is based on OpenHermes 2.5 - Mistral 7B and fine-tuned to emulate the target behavior of Anthropic’s “Golden Gate Claude” (https://www.anthropic.com/news/golden-gate-claude)(White Paper: https://transformer-circuits.pub/2024/scaling-monosemanticity/index.html)
|
| 7 |
+
This model is the product of a novel fine-tuning technique that enables a model to rapidly generalize data through smaller samples. It is instructional in that it can be used to prompt a model to exhibit a target behavior, similar to a system prompt. In contrast to a system prompt, the behaviors are unaffected by their position in the context, therefore all responses will be Golden Gate Bridge theme even when the queries are not directly relevant to the subject matter.
|
| 8 |
+
|
| 9 |
+
Intended Use:
|
| 10 |
+
This model was created for demonstration purposes.
|
| 11 |
+
|
| 12 |
+
License:
|
| 13 |
+
MIT License
|
| 14 |
+
|
| 15 |
+
The Attentio Team:
|
| 16 |
+
Lukas Valine, Dan Belling, Jacob Brindley, and Julian Besonen
|