five

open-llm-leaderboard-old/details_ankhamun__xxxI-Ixxx

收藏
Hugging Face2024-02-09 更新2024-06-22 收录
下载链接:
https://hf-mirror.com/datasets/open-llm-leaderboard-old/details_ankhamun__xxxI-Ixxx
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ankhamun/xxxI-Ixxx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。数据集是从1次运行中生成的,每次运行在每个配置中表示为特定的分割,分割名称使用运行的时间戳。train分割始终指向最新的结果。一个名为results的额外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。README还提供了如何使用Python中的datasets库加载运行细节的示例。

该数据集是在Open LLM Leaderboard上对模型ankhamun/xxxI-Ixxx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。数据集是从1次运行中生成的,每次运行在每个配置中表示为特定的分割,分割名称使用运行的时间戳。train分割始终指向最新的结果。一个名为results的额外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。README还提供了如何使用Python中的datasets库加载运行细节的示例。
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 ankhamun/xxxI-Ixxx 的评估运行期间自动创建的,用于 Open LLM Leaderboard

数据集组成

  • 数据集包含 63 个配置,每个配置对应一个评估任务。
  • 数据集从 1 次运行中创建。每次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。
  • "train" 分割始终指向最新的结果。
  • 额外的 "results" 配置存储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。

加载数据集示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_ankhamun__xxxI-Ixxx", "harness_winogrande_5", split="train")

最新结果

以下是 2024-02-09T19:25:58.917913 运行的最新结果

python { "all": { "acc": 0.5185710776579808, "acc_stderr": 0.034251914485577906, "acc_norm": 0.5240726925248631, "acc_norm_stderr": 0.03498635392452543, "mc1": 0.3806609547123623, "mc1_stderr": 0.016997627871907926, "mc2": 0.5442191956457653, "mc2_stderr": 0.01519663174796153 }, "harness|arc:challenge|25": { "acc": 0.4931740614334471, "acc_stderr": 0.014610029151379813, "acc_norm": 0.5418088737201365, "acc_norm_stderr": 0.014560220308714695 }, "harness|hellaswag|10": { "acc": 0.5446126269667397, "acc_stderr": 0.004969879532843072, "acc_norm": 0.7254530969926309, "acc_norm_stderr": 0.00445373590094783 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.32, "acc_stderr": 0.04688261722621503, "acc_norm": 0.32, "acc_norm_stderr": 0.04688261722621503 }, # 其他任务的结果省略 }

配置详情

  • config_name: harness_arc_challenge_25

    • 分割: 2024_02_09T19_25_58.917913
      • 路径: **/details_harness|arc:challenge|25_2024-02-09T19-25-58.917913.parquet
    • 分割: latest
      • 路径: **/details_harness|arc:challenge|25_2024-02-09T19-25-58.917913.parquet
  • config_name: harness_gsm8k_5

    • 分割: 2024_02_09T19_25_58.917913
      • 路径: **/details_harness|gsm8k|5_2024-02-09T19-25-58.917913.parquet
    • 分割: latest
      • 路径: **/details_harness|gsm8k|5_2024-02-09T19-25-58.917913.parquet
  • config_name: harness_hellaswag_10

    • 分割: 2024_02_09T19_25_58.917913
      • 路径: **/details_harness|hellaswag|10_2024-02-09T19-25-58.917913.parquet
    • 分割: latest
      • 路径: **/details_harness|hellaswag|10_2024-02-09T19-25-58.917913.parquet
  • config_name: harness_hendrycksTest_5

    • 分割: 2024_02_09T19_25_58.917913
      • 路径:
        • **/details_harness|hendrycksTest-abstract_algebra|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-anatomy|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-astronomy|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-business_ethics|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-clinical_knowledge|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-college_biology|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-college_chemistry|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-college_computer_science|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-college_mathematics|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-college_medicine|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-college_physics|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-computer_security|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-conceptual_physics|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-econometrics|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-electrical_engineering|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-elementary_mathematics|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-formal_logic|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-global_facts|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-high_school_biology|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-high_school_chemistry|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-high_school_computer_science|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-high_school_european_history|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-high_school_geography|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-high_school_mathematics|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-high_school_microeconomics|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-high_school_physics|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-high_school_psychology|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-high_school_statistics|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-high_school_us_history|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-high_school_world_history|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-human_aging|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-human_sexuality|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-international_law|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-jurisprudence|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-logical_fallacies|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-machine_learning|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-management|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-marketing|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-medical_genetics|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-miscellaneous|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-moral_disputes|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-moral_scenarios|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-nutrition|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-philosophy|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-prehistory|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-professional_accounting|5_2024-02-09T19-25-58.917913.parquet
        • **/details_harness|hendrycksTest-professional_law|5_2024-02-09T19-25-58.917913.parquet
        • `**/details_har
二维码
社区交流群
二维码
科研交流群
商业服务