metadata
license: mit
library_name: transformers
MyAwesomeModel
Evaluation Results (step_500 checkpoint)
Scores are reported to three decimal places.
| Benchmark | Score |
|---|---|
| math_reasoning | 0.900 |
| logical_reasoning | 0.920 |
| common_sense | 0.880 |
| reading_comprehension | 0.850 |
| question_answering | 0.800 |
| text_classification | 0.900 |
| sentiment_analysis | 0.880 |
| code_generation | 0.820 |
| creative_writing | 0.780 |
| dialogue_generation | 0.830 |
| summarization | 0.890 |
| translation | 0.930 |
| knowledge_retrieval | 0.850 |
| instruction_following | 0.880 |
| safety_evaluation | 0.870 |
| Overall (weighted) | 0.866 |
Usage
- Recommended temperature: 0.6
- System prompt supported; no special output tokens required.