Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -1,3 +1,14 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
# Steering Vectors for LLM Behavior Control
|
| 2 |
|
| 3 |
Pre-extracted steering vectors for use with [rotalabs-steer](https://github.com/rotalabs/rotalabs-steer).
|
|
@@ -18,12 +29,10 @@ from rotalabs_steer import SteeringVector, ActivationInjector
|
|
| 18 |
vector_path = hf_hub_download(
|
| 19 |
repo_id="rotalabs/steering-vectors",
|
| 20 |
filename="refusal_qwen3_8b/layer_15.pt",
|
| 21 |
-
repo_type="model",
|
| 22 |
)
|
| 23 |
metadata_path = hf_hub_download(
|
| 24 |
repo_id="rotalabs/steering-vectors",
|
| 25 |
filename="refusal_qwen3_8b/layer_15.json",
|
| 26 |
-
repo_type="model",
|
| 27 |
)
|
| 28 |
|
| 29 |
# Load and use
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: mit
|
| 3 |
+
tags:
|
| 4 |
+
- steering-vectors
|
| 5 |
+
- activation-steering
|
| 6 |
+
- llm-safety
|
| 7 |
+
- representation-engineering
|
| 8 |
+
- interpretability
|
| 9 |
+
library_name: rotalabs-steer
|
| 10 |
+
---
|
| 11 |
+
|
| 12 |
# Steering Vectors for LLM Behavior Control
|
| 13 |
|
| 14 |
Pre-extracted steering vectors for use with [rotalabs-steer](https://github.com/rotalabs/rotalabs-steer).
|
|
|
|
| 29 |
vector_path = hf_hub_download(
|
| 30 |
repo_id="rotalabs/steering-vectors",
|
| 31 |
filename="refusal_qwen3_8b/layer_15.pt",
|
|
|
|
| 32 |
)
|
| 33 |
metadata_path = hf_hub_download(
|
| 34 |
repo_id="rotalabs/steering-vectors",
|
| 35 |
filename="refusal_qwen3_8b/layer_15.json",
|
|
|
|
| 36 |
)
|
| 37 |
|
| 38 |
# Load and use
|