open-llm-leaderboard-old/details_AA051611__A0119
收藏Hugging Face2024-01-18 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_AA051611__A0119
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AA051611/A0119进行评估时自动生成的。数据集包含63个配置,每个配置对应一个评估任务。数据集由一个或多个运行结果组成,每个运行结果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。此外,还有一个results配置,用于存储所有运行结果的聚合数据,这些数据用于计算和显示Open LLM Leaderboard上的聚合指标。README还提供了如何使用Hugging Face datasets库加载数据集的示例,并包含了特定运行的最新结果。
该数据集是在Open LLM Leaderboard上对模型AA051611/A0119进行评估时自动生成的。数据集包含63个配置,每个配置对应一个评估任务。数据集由一个或多个运行结果组成,每个运行结果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。此外,还有一个results配置,用于存储所有运行结果的聚合数据,这些数据用于计算和显示Open LLM Leaderboard上的聚合指标。README还提供了如何使用Hugging Face datasets库加载数据集的示例,并包含了特定运行的最新结果。
提供机构:
open-llm-leaderboard-old
原始信息汇总
数据集概述
数据集创建
- 创建背景:该数据集是在模型AA051611/A0119在Open LLM Leaderboard上的评估运行期间自动创建的。
- 数据集组成:包含63个配置,每个配置对应一个评估任务。
- 创建次数:数据集从1次运行中创建,每个运行在每个配置中作为一个特定的分片存在,分片名称使用运行的时间戳。"train"分片始终指向最新的结果。
数据集加载
- 加载示例: python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_AA051611__A0119", "harness_winogrande_5", split="train")
最新结果
- 结果来源:这些是最新结果,来自2024-01-18T23:36:49.726923的运行。
- 结果示例: python { "all": { "acc": 0.7265379913533245, "acc_stderr": 0.029728940123723756, "acc_norm": 0.7314556841951575, "acc_norm_stderr": 0.030296550379322322, "mc1": 0.397796817625459, "mc1_stderr": 0.017133934248559635, "mc2": 0.5795530367826707, "mc2_stderr": 0.015322289394733637 }, "harness|arc:challenge|25": { "acc": 0.6040955631399317, "acc_stderr": 0.014291228393536588, "acc_norm": 0.6424914675767918, "acc_norm_stderr": 0.014005494275916573 }, "harness|hellaswag|10": { "acc": 0.6603266281617207, "acc_stderr": 0.004726304225137318, "acc_norm": 0.8474407488548098, "acc_norm_stderr": 0.0035882728748524817 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.39, "acc_stderr": 0.04902071300001975, "acc_norm": 0.39, "acc_norm_stderr": 0.04902071300001975 }, # 其他任务的结果... }
配置详情
-
配置名称:harness_arc_challenge_25
- 数据文件:
- 分片:2024_01_18T23_36_49.726923
- 路径:
**/details_harness|arc:challenge|25_2024-01-18T23-36-49.726923.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|arc:challenge|25_2024-01-18T23-36-49.726923.parquet
- 路径:
- 分片:2024_01_18T23_36_49.726923
- 数据文件:
-
配置名称:harness_gsm8k_5
- 数据文件:
- 分片:2024_01_18T23_36_49.726923
- 路径:
**/details_harness|gsm8k|5_2024-01-18T23-36-49.726923.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|gsm8k|5_2024-01-18T23-36-49.726923.parquet
- 路径:
- 分片:2024_01_18T23_36_49.726923
- 数据文件:
-
配置名称:harness_hellaswag_10
- 数据文件:
- 分片:2024_01_18T23_36_49.726923
- 路径:
**/details_harness|hellaswag|10_2024-01-18T23-36-49.726923.parquet
- 路径:
- 分片:latest
- 路径:
**/details_harness|hellaswag|10_2024-01-18T23-36-49.726923.parquet
- 路径:
- 分片:2024_01_18T23_36_49.726923
- 数据文件:
-
配置名称:harness_hendrycksTest_5
- 数据文件:
- 分片:2024_01_18T23_36_49.726923
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-anatomy|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-astronomy|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-business_ethics|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-college_biology|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-college_chemistry|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-college_computer_science|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-college_mathematics|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-college_medicine|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-college_physics|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-computer_security|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-econometrics|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-formal_logic|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-global_facts|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-high_school_biology|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-high_school_chemistry|5_2024-01-18T23-36-49.726923.parquet
其他任务的路径...
- 路径:
- 分片:latest
- 路径:
**/details_harness|hendrycksTest-abstract_algebra|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-anatomy|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-astronomy|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-business_ethics|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-clinical_knowledge|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-college_biology|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-college_chemistry|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-college_computer_science|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-college_mathematics|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-college_medicine|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-college_physics|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-computer_security|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-conceptual_physics|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-econometrics|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-electrical_engineering|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-elementary_mathematics|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-formal_logic|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-global_facts|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-high_school_biology|5_2024-01-18T23-36-49.726923.parquet**/details_harness|hendrycksTest-high_school_chemistry|5_2024-01-18T23-36-49.726923.parquet
其他任务的路径...
- 路径:
- 分片:2024_01_18T23_36_49.726923
- 数据文件:



