five

open-llm-leaderboard-old/details_IkariDev__Athena-tmp

收藏
Hugging Face2023-08-29 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_IkariDev__Athena-tmp
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IkariDev/Athena-tmp时自动创建的,用于Open LLM Leaderboard的评估。数据集包含61个配置,每个配置对应一个评估任务。数据集由1次运行生成,每次运行的结果存储为特定配置中的一个分割,分割名称使用运行的时间戳。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,用于计算和显示Open LLM Leaderboard上的聚合指标。

该数据集是在评估模型IkariDev/Athena-tmp时自动创建的,用于Open LLM Leaderboard的评估。数据集包含61个配置,每个配置对应一个评估任务。数据集由1次运行生成,每次运行的结果存储为特定配置中的一个分割,分割名称使用运行的时间戳。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,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

该数据集是在模型 IkariDev/Athena-tmpOpen LLM Leaderboard 上的评估运行期间自动创建的。数据集包含 61 个配置,每个配置对应一个评估任务。

数据集结构

  • 配置数量:61 个配置
  • 数据来源:1 次运行,每个运行可以在每个配置中找到特定的分割,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。
  • 聚合结果:一个额外的配置 "results" 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_IkariDev__Athena-tmp", "harness_truthfulqa_mc_0", split="train")

最新结果

以下是 2023-08-29T15:50:42.106753 运行的最新结果

python { "all": { "acc": 0.5888874553745688, "acc_stderr": 0.03407664559390293, "acc_norm": 0.5926858740874733, "acc_norm_stderr": 0.034057449595187576, "mc1": 0.38922888616891066, "mc1_stderr": 0.017068552680690328, "mc2": 0.5536706803409501, "mc2_stderr": 0.01611557269809252 }, "harness|arc:challenge|25": { "acc": 0.5674061433447098, "acc_stderr": 0.014478005694182531, "acc_norm": 0.5921501706484642, "acc_norm_stderr": 0.014361097288449696 }, "harness|hellaswag|10": { "acc": 0.6218880701055567, "acc_stderr": 0.004839247332606038, "acc_norm": 0.8212507468631747, "acc_norm_stderr": 0.003823591814133031 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.29, "acc_stderr": 0.045604802157206845, "acc_norm": 0.29, "acc_norm_stderr": 0.045604802157206845 }, ... }

配置详情

  • 配置名称:harness_arc_challenge_25

    • 数据文件
      • 分割:2023_08_29T15_50_42.106753
        • 路径:**/details_harness|arc:challenge|25_2023-08-29T15:50:42.106753.parquet
      • 分割:latest
        • 路径:**/details_harness|arc:challenge|25_2023-08-29T15:50:42.106753.parquet
  • 配置名称:harness_hellaswag_10

    • 数据文件
      • 分割:2023_08_29T15_50_42.106753
        • 路径:**/details_harness|hellaswag|10_2023-08-29T15:50:42.106753.parquet
      • 分割:latest
        • 路径:**/details_harness|hellaswag|10_2023-08-29T15:50:42.106753.parquet
  • 配置名称:harness_hendrycksTest

    • 数据文件
      • 分割:2023_08_29T15_50_42.106753
        • 路径:
          • **/details_harness|hendrycksTest-abstract_algebra|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-anatomy|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-astronomy|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-business_ethics|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-college_biology|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-college_chemistry|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-college_computer_science|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-college_mathematics|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-college_medicine|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-college_physics|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-computer_security|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-conceptual_physics|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-econometrics|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-electrical_engineering|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-formal_logic|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-global_facts|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-high_school_biology|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-high_school_european_history|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-high_school_geography|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-high_school_physics|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-high_school_psychology|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-high_school_statistics|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-high_school_us_history|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-high_school_world_history|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-human_aging|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-human_sexuality|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-international_law|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-jurisprudence|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-logical_fallacies|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-machine_learning|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-management|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-marketing|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-medical_genetics|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-miscellaneous|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-moral_disputes|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-moral_scenarios|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-nutrition|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-philosophy|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-prehistory|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-professional_accounting|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-professional_law|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-professional_medicine|5_2023-08-29T15:50:42.106753.parquet
          • **/details_harness|hendrycksTest-professional_psychology|5_2023-08-29T15:50:42.106753.parquet
          • `**/details_harness|hendrycksTest-public_relations|5_2
5,000+
优质数据集
54 个
任务类型
进入经典数据集
二维码
社区交流群

面向社区/商业的数据集话题

二维码
科研交流群

面向高校/科研机构的开源数据集话题

数据驱动未来

携手共赢发展

商业合作