Text Generation
PEFT
Safetensors
lora
topk-lora
gradient-routing
sleeper-agent
backdoor
interpretability
ai-safety
gemma2
Instructions to use interpretable-finetuning/gradient-routing-gemma with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- PEFT
How to use interpretable-finetuning/gradient-routing-gemma with PEFT:
Task type is invalid.
- Notebooks
- Google Colab
- Kaggle
Card: link interpretable-finetuning/topklora -> interpretable-finetuning/topklora-gemma-2-2b (repo renamed)
Browse files
README.md
CHANGED
|
@@ -28,7 +28,7 @@ scored against it rather than only against behaviour.
|
|
| 28 |
|
| 29 |
The companion release of normally trained adapters (3 layer families × 5 seeds, with their
|
| 30 |
discovered circuits) is
|
| 31 |
-
[`interpretable-finetuning/topklora`](https://huggingface.co/interpretable-finetuning/topklora).
|
| 32 |
The `unrouted/` twins here are the same recipe as its `l1523` family.
|
| 33 |
|
| 34 |
---
|
|
|
|
| 28 |
|
| 29 |
The companion release of normally trained adapters (3 layer families × 5 seeds, with their
|
| 30 |
discovered circuits) is
|
| 31 |
+
[`interpretable-finetuning/topklora-gemma-2-2b`](https://huggingface.co/interpretable-finetuning/topklora-gemma-2-2b).
|
| 32 |
The `unrouted/` twins here are the same recipe as its `l1523` family.
|
| 33 |
|
| 34 |
---
|