| language: | |
| - en | |
| base_model: | |
| - Qwen/QwQ-32B | |
| Trained with a small, concise dataset that does "less thinking". | |
| In my limited experiments it does think less while matching the output quality. | |
| **TRY this system message:** | |
| You are a thinking assistant. Think in a concise and efficient way. |