open-llm-leaderboard-old/details_Qwen__Qwen1.5-4B
收藏Hugging Face2024-02-18 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_Qwen__Qwen1.5-4B
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是在评估模型Qwen/Qwen1.5-4B在Open LLM Leaderboard上的运行过程中自动创建的。数据集由63个配置组成,每个配置对应一个评估任务。数据集由1次运行生成,每次运行可以在每个配置的特定分割中找到,分割以运行的时间戳命名。train分割始终指向最新结果。此外,results配置存储了所有运行的聚合结果,并用于计算和显示在Open LLM Leaderboard上的聚合指标。
该数据集是在评估模型Qwen/Qwen1.5-4B在Open LLM Leaderboard上的运行过程中自动创建的。数据集由63个配置组成,每个配置对应一个评估任务。数据集由1次运行生成,每次运行可以在每个配置的特定分割中找到,分割以运行的时间戳命名。train分割始终指向最新结果。此外,results配置存储了所有运行的聚合结果,并用于计算和显示在Open LLM Leaderboard上的聚合指标。
提供机构:
open-llm-leaderboard-old
原始信息汇总
数据集概述
数据集名称
- 名称: Evaluation run of Qwen/Qwen1.5-4B
数据集组成
- 配置数量: 63个配置
- 每个配置对应: 一个评估任务
- 数据来源: 从1次运行中创建
- 分割命名: 使用运行的时间戳
- 训练分割: 指向最新结果
额外配置
- 结果配置: 存储所有运行的聚合结果,用于计算和显示在Open LLM Leaderboard上的聚合指标
数据加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_Qwen__Qwen1.5-4B", "harness_winogrande_5", split="train")
最新结果
- 最新结果来源: 2024-02-18T06:08:09.461574
- 结果示例:
python
{
"all": {
"acc": 0.563518335394339,
"acc_stderr": 0.03397274030232774,
"acc_norm": 0.5664210423658909,
"acc_norm_stderr": 0.03465890381962306,
"mc1": 0.31701346389228885,
"mc1_stderr": 0.016289203374403392,
"mc2": 0.4727318979180884,
"mc2_stderr": 0.014620457003931853
},
"harness|arc:challenge|25": {
"acc": 0.44795221843003413,
"acc_stderr": 0.01453201149821167,
"acc_norm": 0.48464163822525597,
"acc_norm_stderr": 0.01460449612939491
},
"harness|hellaswag|10": {
"acc": 0.5237004580760805,
"acc_stderr": 0.004984172621822883,
"acc_norm": 0.7157936666002789,
"acc_norm_stderr": 0.004501137895230714
},
更多任务结果...
}
配置详情
-
配置名称: harness_arc_challenge_25
- 数据文件:
- 分割: 2024_02_18T06_08_09.461574
- 路径: **/details_harness|arc:challenge|25_2024-02-18T06-08-09.461574.parquet
- 分割: latest
- 路径: **/details_harness|arc:challenge|25_2024-02-18T06-08-09.461574.parquet
- 分割: 2024_02_18T06_08_09.461574
- 数据文件:
-
配置名称: harness_gsm8k_5
- 数据文件:
- 分割: 2024_02_18T06_08_09.461574
- 路径: **/details_harness|gsm8k|5_2024-02-18T06-08-09.461574.parquet
- 分割: latest
- 路径: **/details_harness|gsm8k|5_2024-02-18T06-08-09.461574.parquet
- 分割: 2024_02_18T06_08_09.461574
- 数据文件:
-
配置名称: harness_hellaswag_10
- 数据文件:
- 分割: 2024_02_18T06_08_09.461574
- 路径: **/details_harness|hellaswag|10_2024-02-18T06-08-09.461574.parquet
- 分割: latest
- 路径: **/details_harness|hellaswag|10_2024-02-18T06-08-09.461574.parquet
- 分割: 2024_02_18T06_08_09.461574
- 数据文件:
-
配置名称: harness_hendrycksTest_5
- 数据文件:
- 分割: 2024_02_18T06_08_09.461574
- 路径:
- **/details_harness|hendrycksTest-abstract_algebra|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-anatomy|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-astronomy|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-business_ethics|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-clinical_knowledge|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-college_biology|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-college_chemistry|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-college_computer_science|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-college_mathematics|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-college_medicine|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-college_physics|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-computer_security|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-conceptual_physics|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-econometrics|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-electrical_engineering|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-elementary_mathematics|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-formal_logic|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-global_facts|5_2024-02-18T06-08-09.461574.parquet
- **/details_harness|hendrycksTest-high_school_biology|5_2024-02-18T06-08-09.461574.parquet
更多路径...
- 路径:
- 分割: 2024_02_18T06_08_09.461574
- 数据文件:



