Instructions to use TokenBender/circuit-discovery with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use TokenBender/circuit-discovery with Transformers:
# Load model directly from transformers import AutoModel model = AutoModel.from_pretrained("TokenBender/circuit-discovery", dtype="auto") - Notebooks
- Google Colab
- Kaggle
Add minimal model card
Browse files
README.md
ADDED
|
@@ -0,0 +1,11 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
library_name: transformers
|
| 3 |
+
base_model: Qwen/Qwen2.5-Math-1.5B
|
| 4 |
+
license: other
|
| 5 |
+
---
|
| 6 |
+
|
| 7 |
+
# circuit-discovery
|
| 8 |
+
|
| 9 |
+
Public checkpoint repository for arithmetic circuit-discovery experiments.
|
| 10 |
+
|
| 11 |
+
Contents are organized as checkpoint artifacts and adapter artifacts. See the source project for experiment context.
|