Model
Phi-3-medium-128k-instruct
Microsoft · open weights
Knowledge & Reasoning
Knowledge & Reasoning| Benchmark | Score | Metric | Harness | Source | Confidence | Link |
|---|---|---|---|---|---|---|
| ANLI | 55.8 | score | N/A | Self-reported | medium | arxiv.org |
| BBH | 81.4 | average | N/A | Self-reported | medium | arxiv.org |
| GPQA Diamond | 27.6 | accuracy | N/A | Independent | high | epoch.ai |
| HellaSwag | 82.4 | overall_accuracy | N/A | Self-reported | medium | arxiv.org |
| MMLU | 78.0 | exact_match | N/A | Self-reported | medium | arxiv.org |
| MMLU | 77.5 | exact_match | N/A | Independent | medium | crfm.stanford.edu |
| OpenBookQA | 87.4 | accuracy | N/A | Self-reported | medium | arxiv.org |
| Trivia QA | 73.9 | exact_match | N/A | Self-reported | medium | arxiv.org |
| WinoGrande | 81.5 | accuracy | N/A | Self-reported | medium | arxiv.org |
Reproduced by the community
Loading community runs…