open-llm-leaderboard-old/details_brucethemoose__Yi-34B-200K-DARE-megamerge-v8
收藏Hugging Face2024-01-15 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_brucethemoose__Yi-34B-200K-DARE-megamerge-v8
下载链接
链接失效反馈官方服务:
资源简介:
该数据集是在模型brucethemoose/Yi-34B-200K-DARE-megamerge-v8的评估运行期间自动创建的,用于在Open LLM Leaderboard上进行评估。数据集由63个配置组成,每个配置对应一个评估任务。数据集是从1次运行中创建的,每次运行可以在每个配置中找到,分割使用运行的时间戳命名。"train"分割始终指向最新的结果。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Open LLM Leaderboard上的聚合指标。
该数据集是在模型brucethemoose/Yi-34B-200K-DARE-megamerge-v8的评估运行期间自动创建的,用于在Open LLM Leaderboard上进行评估。数据集由63个配置组成,每个配置对应一个评估任务。数据集是从1次运行中创建的,每次运行可以在每个配置中找到,分割使用运行的时间戳命名。"train"分割始终指向最新的结果。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 brucethemoose/Yi-34B-200K-DARE-megamerge-v8
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。
- 创建来源: 数据集由1次运行创建,每个运行结果存储在特定的分割中,分割名称使用运行的时间戳。
- 最新结果: "train"分割始终指向最新结果。
- 汇总结果: 一个额外的配置"results"存储所有运行的汇总结果,用于计算和显示在Open LLM Leaderboard上的汇总指标。
数据加载示例
python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_brucethemoose__Yi-34B-200K-DARE-megamerge-v8", "harness_winogrande_5", split="train")
最新结果
- 最新结果时间戳: 2024-01-15T22:55:33.545655
- 详细结果: 包含多个任务的评估结果,例如:
- harness|arc:challenge|25:
acc: 0.6501706484641638acc_stderr: 0.013936809212158296acc_norm: 0.6774744027303754acc_norm_stderr: 0.013659980894277371
- harness|hellaswag|10:
acc: 0.6590320653256323acc_stderr: 0.004730658073041563acc_norm: 0.8605855407289384acc_norm_stderr: 0.0034567060380547555
- harness|hendrycksTest-abstract_algebra|5:
acc: 0.47acc_stderr: 0.05016135580465919acc_norm: 0.47acc_norm_stderr: 0.05016135580465919
- 其他任务结果: 类似格式,包含多个任务的评估结果。
- harness|arc:challenge|25:
配置详情
- 配置名称: harness_arc_challenge_25
- 数据文件:
- 分割: 2024_01_15T22_55_33.545655
- 路径: **/details_harness|arc:challenge|25_2024-01-15T22-55-33.545655.parquet
- 分割: latest
- 路径: **/details_harness|arc:challenge|25_2024-01-15T22-55-33.545655.parquet
- 分割: 2024_01_15T22_55_33.545655
- 数据文件:
- 配置名称: harness_gsm8k_5
- 数据文件:
- 分割: 2024_01_15T22_55_33.545655
- 路径: **/details_harness|gsm8k|5_2024-01-15T22-55-33.545655.parquet
- 分割: latest
- 路径: **/details_harness|gsm8k|5_2024-01-15T22-55-33.545655.parquet
- 分割: 2024_01_15T22_55_33.545655
- 数据文件:
- 配置名称: harness_hellaswag_10
- 数据文件:
- 分割: 2024_01_15T22_55_33.545655
- 路径: **/details_harness|hellaswag|10_2024-01-15T22-55-33.545655.parquet
- 分割: latest
- 路径: **/details_harness|hellaswag|10_2024-01-15T22-55-33.545655.parquet
- 分割: 2024_01_15T22_55_33.545655
- 数据文件:
- 配置名称: harness_hendrycksTest_5
- 数据文件:
- 分割: 2024_01_15T22_55_33.545655
- 路径:
- **/details_harness|hendrycksTest-abstract_algebra|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-anatomy|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-astronomy|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-business_ethics|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-clinical_knowledge|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-college_biology|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-college_chemistry|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-college_computer_science|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-college_mathematics|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-college_medicine|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-college_physics|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-computer_security|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-conceptual_physics|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-econometrics|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-electrical_engineering|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-elementary_mathematics|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-formal_logic|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-global_facts|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-high_school_biology|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-high_school_chemistry|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-high_school_computer_science|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-high_school_european_history|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-high_school_geography|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-high_school_government_and_politics|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-high_school_macroeconomics|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-high_school_mathematics|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-high_school_microeconomics|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-high_school_physics|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-high_school_psychology|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-high_school_statistics|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-high_school_us_history|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-high_school_world_history|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-human_aging|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-human_sexuality|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-international_law|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-jurisprudence|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-logical_fallacies|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-machine_learning|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-management|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-marketing|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-medical_genetics|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-miscellaneous|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-moral_disputes|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-moral_scenarios|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-nutrition|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-philosophy|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-prehistory|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-professional_accounting|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-professional_law|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-professional_medicine|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-professional_psychology|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-public_relations|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-security_studies|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-sociology|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hendrycksTest-us_foreign_policy|5_2024-01-15T22-55-33.545655.parquet
- **/details_harness|hend
- 路径:
- 分割: 2024_01_15T22_55_33.545655
- 数据文件:
搜集汇总
背景与挑战
背景概述
该数据集是用于评估模型brucethemoose/Yi-34B-200K-DARE-megamerge-v8在Open LLM Leaderboard上的性能的自动生成数据集,包含63个评估任务配置,并存储了所有运行的聚合结果以计算和显示排行榜指标。
以上内容由遇见数据集搜集并总结生成



