测试版
排行榜/Qwen2 7B Instruct
Alibaba logo

Qwen2 7B Instruct

开源

来自 Alibaba · 发布于 2024-06-04

50.5
平均分
N/A
输入价格
N/A
输出价格
N/A
上下文窗口
text-generation
类型

Tested on 25 benchmarks with 50.5% average. Top scores: JSQuAD (89.6%), JCommonsenseQA (89.1%), JNLI (81.3%).

基准测试类别分数Bar
JSQuADlanguage89.6
JCommonsenseQAlanguage89.1
JNLIlanguage81.3
MMMLU — Chineselanguage61.8
MMMLU — Frenchlanguage60.8
MMMLU — Spanishlanguage60.2
MMMLU — Portugueselanguage60.1
MMMLU — Italianlanguage59.0
MMMLU — Germanlanguage57.1
IFEvallanguage56.8
MMMLU — Japaneselanguage56.6
JMMLUlanguage56.5
MMMLU — Indonesianlanguage54.1
MMMLU — Koreanlanguage54.0
LLM-JP — Overalllanguage51.7
MMMLU — Arabiclanguage50.7
MMMLU — Hindilanguage45.1
MMMLU — Bengalilanguage43.4
BBH (HuggingFace)general37.8
MMMLU — Swahililanguage34.3
MMLU-PROknowledge31.6
MMMLU — Yorubalanguage30.2
MATH Level 5math27.6
MUSRreasoning7.4
GPQAknowledge6.4