open-llm-leaderboard-old/details_Devio__testC

该数据集是在模型Devio/testC的评估运行期间自动创建的，用于Open LLM Leaderboard。数据集由61个配置组成，每个配置对应一个评估任务。数据集从1次运行中创建，每次运行可以在每个配置的特定分割中找到，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，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

数据集来源

该数据集是在对模型 Devio/testC 进行评估运行期间自动创建的，评估运行在 Open LLM Leaderboard 上进行。

数据集结构

• 配置数量：61个配置，每个配置对应一个评估任务。
• 创建次数：从1次运行中创建。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。
• 训练分割："train" 分割始终指向最新的结果。
• 结果配置：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 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_Devio__testC", "harness_truthfulqa_mc_0", split="train")

最新结果

以下是 2023-09-02T17:27:16.860385 运行 的最新结果：

python { "all": { "acc": 0.28185588236286707, "acc_stderr": 0.03225753349873974, "acc_norm": 0.2855290591736718, "acc_norm_stderr": 0.03226027924923892, "mc1": 0.20318237454100369, "mc1_stderr": 0.014085666526340882, "mc2": 0.35665813452391837, "mc2_stderr": 0.014271431688144938 }, "harness|arc:challenge|25": { "acc": 0.35494880546075086, "acc_stderr": 0.013983036904094097, "acc_norm": 0.39590443686006827, "acc_norm_stderr": 0.014291228393536583 }, "harness|hellaswag|10": { "acc": 0.4529974108743278, "acc_stderr": 0.004967685204073108, "acc_norm": 0.6287592113124876, "acc_norm_stderr": 0.004821492994082116 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.3, "acc_stderr": 0.04605661864718381, "acc_norm": 0.3, "acc_norm_stderr": 0.04605661864718381 }, # 其他任务的结果... "harness|truthfulqa:mc|0": { "mc1": 0.20318237454100369, "mc1_stderr": 0.014085666526340882, "mc2": 0.35665813452391837, "mc2_stderr": 0.014271431688144938 } }

配置详情

• 配置名称：harness_arc_challenge_25

  • 数据文件：
    • 分割：2023_09_02T17_27_16.860385
      • 路径：**/details_harness|arc:challenge|25_2023-09-02T17:27:16.860385.parquet
    • 分割：latest
      • 路径：**/details_harness|arc:challenge|25_2023-09-02T17:27:16.860385.parquet

• 配置名称：harness_hellaswag_10

  • 数据文件：
    • 分割：2023_09_02T17_27_16.860385
      • 路径：**/details_harness|hellaswag|10_2023-09-02T17:27:16.860385.parquet
    • 分割：latest
      • 路径：**/details_harness|hellaswag|10_2023-09-02T17:27:16.860385.parquet

• 配置名称：harness_hendrycksTest_5

  • 数据文件：
    • 分割：2023_09_02T17_27_16.860385
      • 路径：
        • **/details_harness|hendrycksTest-abstract_algebra|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-anatomy|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-astronomy|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-business_ethics|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-college_biology|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-college_chemistry|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-college_computer_science|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-college_mathematics|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-college_medicine|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-college_physics|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-computer_security|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-conceptual_physics|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-econometrics|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-electrical_engineering|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-formal_logic|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-global_facts|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-high_school_biology|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-high_school_european_history|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-high_school_geography|5_2023-09-02T17:27:16.860385.parquet
        • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-09-02T17:27:16.860385.parquet

        其他任务的路径...

}