open-llm-leaderboard/details_Sao10K__Zephyrus-L1-33B

该数据集是在模型Sao10K/Zephyrus-L1-33B在Open LLM Leaderboard上的评估运行期间自动创建的。数据集由64个配置组成，每个配置对应一个被评估的任务。数据集由2次运行创建，每次运行在每个配置中表示为特定的分割。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，这些结果用于计算和显示Open LLM Leaderboard上的聚合指标。README文件还提供了如何使用`datasets`库中的`load_dataset`函数加载运行中的详细信息的示例。README中还包含了特定运行的最新结果。

原始信息汇总

数据集概述

数据集简介

该数据集是在对模型 Sao10K/Zephyrus-L1-33B 进行评估运行期间自动创建的，用于 Open LLM Leaderboard。

数据集结构

• 配置数量：64个配置，每个配置对应一个评估任务。
• 运行次数：数据集从2次运行中创建。每个运行在每个配置中作为一个特定的分割存在，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，用于计算和显示聚合指标在 Open LLM Leaderboard 上。

数据加载示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_Sao10K__Zephyrus-L1-33B", "harness_winogrande_5", split="train")

最新结果

以下是 2023-10-26T22:58:44.962753 运行的最新结果： python { "all": { "em": 0.1826761744966443, "em_stderr": 0.003957106009101816, "f1": 0.2689282718120811, "f1_stderr": 0.004022541310882054, "acc": 0.5188394618630148, "acc_stderr": 0.011447189197576924 }, "harness|drop|3": { "em": 0.1826761744966443, "em_stderr": 0.003957106009101816, "f1": 0.2689282718120811, "f1_stderr": 0.004022541310882054 }, "harness|gsm8k|5": { "acc": 0.2357846853677028, "acc_stderr": 0.0116925156506668 }, "harness|winogrande|5": { "acc": 0.8018942383583267, "acc_stderr": 0.011201862744487047 } }

配置详情

以下是部分配置的详细信息：

• 配置名称：harness_arc_challenge_25

  • 数据文件：
    • 分割：2023_10_04T01_26_11.799613
      • 路径：**/details_harness|arc:challenge|25_2023-10-04T01-26-11.799613.parquet
    • 分割：latest
      • 路径：**/details_harness|arc:challenge|25_2023-10-04T01-26-11.799613.parquet

• 配置名称：harness_drop_3

  • 数据文件：
    • 分割：2023_10_26T22_58_44.962753
      • 路径：**/details_harness|drop|3_2023-10-26T22-58-44.962753.parquet
    • 分割：latest
      • 路径：**/details_harness|drop|3_2023-10-26T22-58-44.962753.parquet

• 配置名称：harness_gsm8k_5

  • 数据文件：
    • 分割：2023_10_26T22_58_44.962753
      • 路径：**/details_harness|gsm8k|5_2023-10-26T22-58-44.962753.parquet
    • 分割：latest
      • 路径：**/details_harness|gsm8k|5_2023-10-26T22-58-44.962753.parquet

• 配置名称：harness_hellaswag_10

  • 数据文件：
    • 分割：2023_10_04T01_26_11.799613
      • 路径：**/details_harness|hellaswag|10_2023-10-04T01-26-11.799613.parquet
    • 分割：latest
      • 路径：**/details_harness|hellaswag|10_2023-10-04T01-26-11.799613.parquet

• 配置名称：harness_hendrycksTest_5

  • 数据文件：
    • 分割：2023_10_04T01_26_11.799613
      • 路径：
        • **/details_harness|hendrycksTest-abstract_algebra|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-anatomy|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-astronomy|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-business_ethics|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-clinical_knowledge|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-college_biology|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-college_chemistry|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-college_computer_science|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-college_mathematics|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-college_medicine|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-college_physics|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-computer_security|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-conceptual_physics|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-econometrics|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-electrical_engineering|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-elementary_mathematics|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-formal_logic|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-global_facts|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-high_school_biology|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-high_school_chemistry|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-high_school_computer_science|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-high_school_european_history|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-high_school_geography|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-high_school_government_and_politics|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-high_school_macroeconomics|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-high_school_mathematics|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-high_school_microeconomics|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-high_school_physics|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-high_school_psychology|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-high_school_statistics|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-high_school_us_history|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-high_school_world_history|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-human_aging|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-human_sexuality|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-international_law|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-jurisprudence|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-logical_fallacies|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-machine_learning|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-management|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-marketing|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-medical_genetics|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-miscellaneous|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-moral_disputes|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-moral_scenarios|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-nutrition|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-philosophy|5_2023-10-04T01-26-11.799613.parquet
        • **/details_harness|hendrycksTest-prehistory|5_2023-10-04T01-26-11.799613.parquet
        • `**/details_harness|hendrycksTest-professional_accounting|5_2023-10-04T01-26-11