open-llm-leaderboard-old/details_UCLA-AGI__test
收藏Hugging Face2024-01-05 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_UCLA-AGI__test
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是在Open LLM Leaderboard上对模型UCLA-AGI/test进行评估运行时自动创建的。数据集由63个配置组成,每个配置对应一个评估任务。数据集包含1次运行的结果,每次运行作为每个配置中的一个特定分割存储。train分割始终指向最新的结果。一个额外的配置results存储了所有运行的聚合结果,用于计算和显示Open LLM Leaderboard上的聚合指标。README文件还提供了如何使用`datasets`库中的`load_dataset`函数加载运行细节的示例。最新的运行结果也包含在README中。
该数据集是在Open LLM Leaderboard上对模型UCLA-AGI/test进行评估运行时自动创建的。数据集由63个配置组成,每个配置对应一个评估任务。数据集包含1次运行的结果,每次运行作为每个配置中的一个特定分割存储。train分割始终指向最新的结果。一个额外的配置results存储了所有运行的聚合结果,用于计算和显示Open LLM Leaderboard上的聚合指标。README文件还提供了如何使用`datasets`库中的`load_dataset`函数加载运行细节的示例。最新的运行结果也包含在README中。
提供机构:
open-llm-leaderboard-old
原始信息汇总
数据集概述
数据集基本信息
- 名称: Evaluation run of UCLA-AGI/test
- 来源: 在模型 UCLA-AGI/test 的评估运行期间自动创建
- 创建时间: 在 Open LLM Leaderboard 上的评估运行期间
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: 来自1次运行,每个运行在每个配置中作为一个特定的分割存在,分割名称使用运行的时间戳
- 训练分割: 总是指向最新的结果
- 额外配置: "results" 存储所有运行的聚合结果,用于计算和显示聚合指标在 Open LLM Leaderboard 上
数据加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_UCLA-AGI__test", "harness_winogrande_5", split="train")
最新结果
- 最新结果时间戳: 2024-01-05T00:36:41.239145
- 结果示例:
python
{
"all": {
"acc": 0.6083070915056247,
"acc_stderr": 0.032946994490981554,
"acc_norm": 0.6144414847119565,
"acc_norm_stderr": 0.03362118464961407,
"mc1": 0.408812729498164,
"mc1_stderr": 0.01720995215164173,
"mc2": 0.5738893811625738,
"mc2_stderr": 0.015990080392547533
},
"harness|arc:challenge|25": {
"acc": 0.6168941979522184,
"acc_stderr": 0.014206472661672876,
"acc_norm": 0.658703071672355,
"acc_norm_stderr": 0.013855831287497728
},
"harness|hellaswag|10": {
"acc": 0.6759609639514041,
"acc_stderr": 0.004670581884781161,
"acc_norm": 0.8544114718183629,
"acc_norm_stderr": 0.003519724163310889
},
其他任务结果省略
}
配置详情
-
配置名称: harness_arc_challenge_25
- 数据文件:
- 分割: 2024_01_05T00_36_41.239145
- 路径: **/details_harness|arc:challenge|25_2024-01-05T00-36-41.239145.parquet
- 分割: latest
- 路径: **/details_harness|arc:challenge|25_2024-01-05T00-36-41.239145.parquet
- 分割: 2024_01_05T00_36_41.239145
- 数据文件:
-
配置名称: harness_gsm8k_5
- 数据文件:
- 分割: 2024_01_05T00_36_41.239145
- 路径: **/details_harness|gsm8k|5_2024-01-05T00-36-41.239145.parquet
- 分割: latest
- 路径: **/details_harness|gsm8k|5_2024-01-05T00-36-41.239145.parquet
- 分割: 2024_01_05T00_36_41.239145
- 数据文件:
-
配置名称: harness_hellaswag_10
- 数据文件:
- 分割: 2024_01_05T00_36_41.239145
- 路径: **/details_harness|hellaswag|10_2024-01-05T00-36-41.239145.parquet
- 分割: latest
- 路径: **/details_harness|hellaswag|10_2024-01-05T00-36-41.239145.parquet
- 分割: 2024_01_05T00_36_41.239145
- 数据文件:
-
配置名称: harness_hendrycksTest_5
- 数据文件:
- 分割: 2024_01_05T00_36_41.239145
- 路径:
- **/details_harness|hendrycksTest-abstract_algebra|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-anatomy|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-astronomy|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-business_ethics|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-clinical_knowledge|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-college_biology|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-college_chemistry|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-college_computer_science|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-college_mathematics|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-college_medicine|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-college_physics|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-computer_security|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-conceptual_physics|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-econometrics|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-electrical_engineering|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-elementary_mathematics|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-formal_logic|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-global_facts|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-high_school_biology|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-high_school_chemistry|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-high_school_computer_science|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-high_school_european_history|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-high_school_geography|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-high_school_government_and_politics|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-high_school_macroeconomics|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-high_school_mathematics|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-high_school_microeconomics|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-high_school_physics|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-high_school_psychology|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-high_school_statistics|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-high_school_us_history|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-high_school_world_history|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-human_aging|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-human_sexuality|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-international_law|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-jurisprudence|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-logical_fallacies|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-machine_learning|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-management|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-marketing|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-medical_genetics|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-miscellaneous|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-moral_disputes|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-moral_scenarios|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-nutrition|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-philosophy|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-prehistory|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-professional_accounting|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-professional_law|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-professional_medicine|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-professional_psychology|5_2024-01-05T00-36-41.239145.parquet
- **/details_harness|hendrycksTest-public_relations|5_2024-01
- 路径:
- 分割: 2024_01_05T00_36_41.239145
- 数据文件:



