Upload README.md with huggingface_hub
Browse files
README.md
CHANGED
|
@@ -37,8 +37,10 @@ Question: "What was OpenAI's revenue in Q2 2026?" — unverifiable, 20 resamples
|
|
| 37 |
| Model | Refusals | Fabrications |
|
| 38 |
|-------|----------|--------------|
|
| 39 |
| **Hermes-3-binary** | **20/20** | **0/20** |
|
| 40 |
-
|
|
| 41 |
| Best old SFT (C) | 13/20 | 8/20 |
|
|
|
|
|
|
|
| 42 |
|
| 43 |
## Usage
|
| 44 |
|
|
|
|
| 37 |
| Model | Refusals | Fabrications |
|
| 38 |
|-------|----------|--------------|
|
| 39 |
| **Hermes-3-binary** | **20/20** | **0/20** |
|
| 40 |
+
| Hermes-3 Base | 9/20 | 11/20 |
|
| 41 |
| Best old SFT (C) | 13/20 | 8/20 |
|
| 42 |
+
| ABCD (4-specialist merge) | 5/20 | 11/20 |
|
| 43 |
+
| AB (dual merge) | 2/20 | 11/20 |
|
| 44 |
|
| 45 |
## Usage
|
| 46 |
|