Spaces:
Running
Running
feifeinoban commited on
Commit ·
13aaaf8
1
Parent(s): 4b270af
Force rebuild
Browse files
README.md
CHANGED
|
@@ -80,7 +80,7 @@ This replaces labor-intensive human labeling with **autonomous, human-aligned re
|
|
| 80 |
- At inference, **add vectors to internal activations** (e.g., Layer 18 of Llama 3.1) to guide behavior.
|
| 81 |
- **No fine-tuning needed**—works on closed-source models like GPT-5.
|
| 82 |
|
| 83 |
-

|
| 84 |
|
| 85 |
> *Figure: The MENTOR framework (from paper). Shell implements this full pipeline.*
|
| 86 |
|
|
|
|
| 80 |
- At inference, **add vectors to internal activations** (e.g., Layer 18 of Llama 3.1) to guide behavior.
|
| 81 |
- **No fine-tuning needed**—works on closed-source models like GPT-5.
|
| 82 |
|
| 83 |
+

|
| 84 |
|
| 85 |
> *Figure: The MENTOR framework (from paper). Shell implements this full pipeline.*
|
| 86 |
|