Phi-2
#232 of 274 scored.
MicrosoftDec 12, 2023Open weights
Capabilities Index
Range 88.6–113.8
107.9
Context
Tokens
—
Input
Per 1M tokens
—
Output
Per 1M tokens
—
Benchmarks
No independent scores yet.
Epoch AI hasn't run Phi-2 yet. Its scores show up here the morning after they do.
Everything else Epoch has for it
- ARC AI267.9%
- OpenBookQA64.8%
- BBH45.9%
- TriviaQA45.2%
- MMLU44.5%
- HellaSwag38.1%
- ANLI13.8%
- Winogrande9.4%
On the index
Around it.
Sources