Commit
•
0302b76
1
Parent(s):
5468344
Push model using huggingface_hub.
Browse files
README.md
CHANGED
|
@@ -6,16 +6,20 @@ tags:
|
|
| 6 |
- arxiv:2409.04185
|
| 7 |
- model_hub_mixin
|
| 8 |
- pytorch_model_hub_mixin
|
| 9 |
-
base_model: EleutherAI/pythia-160m-deduped
|
| 10 |
---
|
| 11 |
|
| 12 |
-
# Model Card for
|
| 13 |
|
| 14 |
A Multi-Layer Sparse Autoencoder (MLSAE) trained on the residual stream activation
|
| 15 |
vectors from [EleutherAI/pythia-160m-deduped](https://huggingface.co/EleutherAI/pythia-160m-deduped) with an
|
| 16 |
expansion factor of R = 64 and sparsity k = 32, over 1 billion
|
| 17 |
tokens from [monology/pile-uncopyrighted](https://huggingface.co/datasets/monology/pile-uncopyrighted).
|
| 18 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 19 |
### Model Sources
|
| 20 |
|
| 21 |
- **Repository:** <https://github.com/tim-lawson/mlsae>
|
|
|
|
| 6 |
- arxiv:2409.04185
|
| 7 |
- model_hub_mixin
|
| 8 |
- pytorch_model_hub_mixin
|
|
|
|
| 9 |
---
|
| 10 |
|
| 11 |
+
# Model Card for tim-lawson/sae-pythia-160m-deduped-x64-k32-tfm-layers-8
|
| 12 |
|
| 13 |
A Multi-Layer Sparse Autoencoder (MLSAE) trained on the residual stream activation
|
| 14 |
vectors from [EleutherAI/pythia-160m-deduped](https://huggingface.co/EleutherAI/pythia-160m-deduped) with an
|
| 15 |
expansion factor of R = 64 and sparsity k = 32, over 1 billion
|
| 16 |
tokens from [monology/pile-uncopyrighted](https://huggingface.co/datasets/monology/pile-uncopyrighted).
|
| 17 |
|
| 18 |
+
|
| 19 |
+
This model is a PyTorch Lightning MLSAETransformer module, which includes the underlying
|
| 20 |
+
transformer.
|
| 21 |
+
|
| 22 |
+
|
| 23 |
### Model Sources
|
| 24 |
|
| 25 |
- **Repository:** <https://github.com/tim-lawson/mlsae>
|