open-llm-leaderboard-old/details_nicholasKluge__Aira-Instruct-1B5

该数据集是在评估模型nicholasKluge/Aira-Instruct-1B5时自动创建的，评估过程在Open LLM Leaderboard上进行。数据集由61个配置组成，每个配置对应一个评估任务。数据集由2次运行生成，每次运行的结果作为特定配置中的一个分割，分割名称使用运行的时间戳命名。train分割始终指向最新的结果。此外，还有一个名为results的配置，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。

原始信息汇总

数据集概述

数据集简介

该数据集是在评估模型 nicholasKluge/Aira-Instruct-1B5 在 Open LLM Leaderboard 上的自动创建的。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
• 运行次数：2次
• 数据分割：每个配置包含多个分割，分割名称使用运行的时间戳。"train" 分割始终指向最新的结果。
• 额外配置："results" 配置存储所有运行的聚合结果，用于计算和显示聚合指标。

数据加载示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_nicholasKluge__Aira-Instruct-1B5", "harness_truthfulqa_mc_0", split="train")

最新结果

最新结果来自2023-08-09T20:50:15.527085的运行，包含多个任务的评估结果。以下是部分结果示例：

python { "all": { "acc": 0.26661052193576085, "acc_stderr": 0.03205800992365353, "acc_norm": 0.26841070974417724, "acc_norm_stderr": 0.03207045668897672, "mc1": 0.23011015911872704, "mc1_stderr": 0.014734557959807763, "mc2": 0.401147792785823, "mc2_stderr": 0.01487516040985077 }, "harness|arc:challenge|25": { "acc": 0.25, "acc_stderr": 0.012653835621466646, "acc_norm": 0.2883959044368601, "acc_norm_stderr": 0.013238394422428162 }, "harness|hellaswag|10": { "acc": 0.35829516032662817, "acc_stderr": 0.004785195049889159, "acc_norm": 0.4261103365863374, "acc_norm_stderr": 0.004934995402995939 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.26, "acc_stderr": 0.04408440022768081, "acc_norm": 0.26, "acc_norm_stderr": 0.04408440022768081 }, # 其他任务的结果... }

配置详情

• 配置名称：harness_arc_challenge_25

  • 数据文件：
    • 分割：2023_08_09T20_49_00.649710
      • 路径：**/details_harness|arc:challenge|25_2023-08-09T20:49:00.649710.parquet
    • 分割：2023_08_09T20_50_15.527085
      • 路径：**/details_harness|arc:challenge|25_2023-08-09T20:50:15.527085.parquet
    • 分割：latest
      • 路径：**/details_harness|arc:challenge|25_2023-08-09T20:50:15.527085.parquet

• 配置名称：harness_hellaswag_10

  • 数据文件：
    • 分割：2023_08_09T20_49_00.649710
      • 路径：**/details_harness|hellaswag|10_2023-08-09T20:49:00.649710.parquet
    • 分割：2023_08_09T20_50_15.527085
      • 路径：**/details_harness|hellaswag|10_2023-08-09T20:50:15.527085.parquet
    • 分割：latest
      • 路径：**/details_harness|hellaswag|10_2023-08-09T20:50:15.527085.parquet

• 配置名称：harness_hendrycksTest_5

  • 数据文件：
    • 分割：2023_08_09T20_49_00.649710
      • 路径：
        • **/details_harness|hendrycksTest-abstract_algebra|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-anatomy|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-astronomy|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-business_ethics|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-college_biology|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-college_chemistry|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-college_computer_science|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-college_mathematics|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-college_medicine|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-college_physics|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-computer_security|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-conceptual_physics|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-econometrics|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-electrical_engineering|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-formal_logic|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-global_facts|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-high_school_biology|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-high_school_european_history|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-high_school_geography|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-high_school_physics|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-high_school_psychology|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-high_school_statistics|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-high_school_us_history|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-high_school_world_history|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-human_aging|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-human_sexuality|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-international_law|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-jurisprudence|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-logical_fallacies|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-machine_learning|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-management|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-marketing|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-medical_genetics|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-miscellaneous|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-moral_disputes|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-moral_scenarios|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-nutrition|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-philosophy|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-prehistory|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-professional_accounting|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-professional_law|5_2023-08-09T20:49:00.649710.parquet
        • **/details_harness|hendrycksTest-professional_medicine|5_2023-08-09T20:49:00.649710.parquet