Instructions to use google/gemma-scope-2b-pt-res with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- SAELens
How to use google/gemma-scope-2b-pt-res with SAELens:
# pip install sae-lens from sae_lens import SAE sae, cfg_dict, sparsity = SAE.from_pretrained( release = "RELEASE_ID", # e.g., "gpt2-small-res-jb". See other options in https://github.com/jbloomAus/SAELens/blob/main/sae_lens/pretrained_saes.yaml sae_id = "SAE_ID", # e.g., "blocks.8.hook_resid_pre". Won't always be a hook point ) - Notebooks
- Google Colab
- Kaggle
Update README.md
#4
by ArthurConmy - opened
README.md
CHANGED
|
@@ -17,7 +17,12 @@ See our [landing page](https://huggingface.co/google/gemma-scope) for details on
|
|
| 17 |
- `2b-pt-`: These SAEs were trained on Gemma v2 2B base model.
|
| 18 |
- `res`: These SAEs were trained on the model's residual stream.
|
| 19 |
|
| 20 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 21 |
|
| 22 |
Point of contact: Arthur Conmy
|
| 23 |
|
|
|
|
| 17 |
- `2b-pt-`: These SAEs were trained on Gemma v2 2B base model.
|
| 18 |
- `res`: These SAEs were trained on the model's residual stream.
|
| 19 |
|
| 20 |
+
|
| 21 |
+
# 3. Which SAE is in the [Neuronpedia demo](https://www.neuronpedia.org/gemma-scope)?
|
| 22 |
+
|
| 23 |
+
https://huggingface.co/google/gemma-scope-2b-pt-res/tree/main/layer_20/width_16k/average_l0_71
|
| 24 |
+
|
| 25 |
+
# 4. Point of Contact
|
| 26 |
|
| 27 |
Point of contact: Arthur Conmy
|
| 28 |
|