docs: point the code link at AlignmentResearch/fibs-probes (paper renamed to Telling FIBS)
Browse files
README.md
CHANGED
|
@@ -23,7 +23,7 @@ base_model:
|
|
| 23 |
Trained activation probes in four architectures (linear, MLP, EFC (early-fusion covariance) and axial)
|
| 24 |
for nine open-weight models. Each probe reads a model's residual-stream activations at six decoder
|
| 25 |
layers and returns one score per transcript. Load them with the `probe-inference` package
|
| 26 |
-
([AlignmentResearch/
|
| 27 |
|
| 28 |
```python
|
| 29 |
from probe_inference import load_probe_from_hub
|
|
|
|
| 23 |
Trained activation probes in four architectures (linear, MLP, EFC (early-fusion covariance) and axial)
|
| 24 |
for nine open-weight models. Each probe reads a model's residual-stream activations at six decoder
|
| 25 |
layers and returns one score per transcript. Load them with the `probe-inference` package
|
| 26 |
+
([AlignmentResearch/fibs-probes](https://github.com/AlignmentResearch/fibs-probes)):
|
| 27 |
|
| 28 |
```python
|
| 29 |
from probe_inference import load_probe_from_hub
|