Benchmarkid
European languages
Mudeli kvaliteet tõlgitud Euroopa keelte võrdlusalustel. EU21 read pärinevad OpenGPT-X hindamisest, mis avaldati koos Teuken 7B Instruct v0.6-ga, mis tõlgib ARC, HellaSwag, TruthfulQA ja MMLU 21-sse kaetud EL-i keelde ja annab iga mudeli kohta ühe keskmise. MMMLU rida on Google'i enda mitmekeelne MMLU näitaja Gemma 4 jaoks.
The EU21 numbers are OpenGPT-X’s, taken from the Teuken 7B Instruct v0.6 model card and the accompanying evaluation preprint; they are machine-translated benchmarks, not native-language test sets. The MMMLU number is Google’s own. LLM EU has run no evaluation and has not verified either harness.
| Mudelid | score | source |
|---|---|---|
| Teuken 7B Instruct v0.6 | 0.57 | link |
| Llama 3.1 8B Instruct | 0.563 | link |
| Mistral 7B Instruct v0.3 | 0.527 | link |
| Aya 23 8B | 0.485 | link |
| Pharia-1-LLM-7B-control-aligned | 0.417 | link |
| Gemma 4 31B IT | 88.4 | link |