Update README.md
Browse files
README.md
CHANGED
|
@@ -46,6 +46,10 @@ Evaluations on the Berkeley Function Calling benchmark demonstrate significant i
|
|
| 46 |
|
| 47 |
These enhancements are due to RL fine-tuning specifically targeted at improving reasoning-based tool calling capabilities.
|
| 48 |
|
|
|
|
|
|
|
|
|
|
|
|
|
| 49 |
## Key Features
|
| 50 |
|
| 51 |
- **Improved Tool Calling in Reasoning Mode**: Reinforcement learning significantly boosts tool usage during complex reasoning tasks.
|
|
|
|
| 46 |
|
| 47 |
These enhancements are due to RL fine-tuning specifically targeted at improving reasoning-based tool calling capabilities.
|
| 48 |
|
| 49 |
+
Eval set accuracy results:
|
| 50 |
+
|
| 51 |
+
|
| 52 |
+
|
| 53 |
## Key Features
|
| 54 |
|
| 55 |
- **Improved Tool Calling in Reasoning Mode**: Reinforcement learning significantly boosts tool usage during complex reasoning tasks.
|