Update README.md
Browse files
README.md
CHANGED
|
@@ -28,8 +28,8 @@ Evaluation was run on a held-out validation set (50 examples):
|
|
| 28 |
|
| 29 |
| Metric | Before FT | After FT |
|
| 30 |
|------|-----------|----------|
|
| 31 |
-
| Tool Selection Accuracy | **
|
| 32 |
-
| Absolute Gain | – | **+
|
| 33 |
|
| 34 |
This shows the model learns **better tool selection and call consistency**, even though the base model already performs strongly.
|
| 35 |
|
|
|
|
| 28 |
|
| 29 |
| Metric | Before FT | After FT |
|
| 30 |
|------|-----------|----------|
|
| 31 |
+
| Tool Selection Accuracy | **88.0%** | **98.0%** |
|
| 32 |
+
| Absolute Gain | – | **+10.0%** |
|
| 33 |
|
| 34 |
This shows the model learns **better tool selection and call consistency**, even though the base model already performs strongly.
|
| 35 |
|