metadata
{}
Introducing SmallThinker-3B-alpha: A Small Model Fine-tuned on QwQ Synthetic Data
We introduce SmallThinker-3B-alpha, a new model fine-tuned from the Qwen2.5-3b-Instruct model using synthetic data generated by QwQ-32B-Preview.
Benchmark Performance
| Model | AIME24 | AMC23 | GAOKAO2024_I | GAOKAO2024_II | MMLU_STEM | AMPS_Hard | math_comp |
|---|---|---|---|---|---|---|---|
| Qwen2.5-3B-Instruct | 6.67 | 45 | 50 | 35.8 | 59.8 | - | - |
| SmallThinker | 16.667 | 57.5 | 64.2 | 57.1 | 68.2 | 70 | 46.8 |
| GPT-4o | 9.3 | - | - | - | 64.2 | 57 | 50 |
Intended Use Cases
SmallThinker is designed for the following use cases:
- Edge Deployment: Its small size makes it ideal for deployment on resource-constrained devices.
- Draft Model for QwQ-32B-Preview: QwQ can serve as a fast and efficient draft model for the larger QwQ-32B-Preview model.
Limitations & Disclaimer
Please be aware of the following limitations:
- Language Limitation: The model has only been trained on English-language datasets, hence its capabilities in other languages are still lacking.
- Unpredictable Outputs: The model may produce unexpected outputs due to its size and probabilistic generation paradigm. Users should exercise caution and validate the model's responses.
- Repetition Issue: The model tends to repeat itself when answering high-difficulty questions. Please increase the
repetition_penaltyto mitigate this issue.