open-llm-leaderboard-old/details_namirocks__mistral-shishya-model-7b-ep3

该数据集是在模型namirocks/mistral-shishya-model-7b-ep3在Open LLM Leaderboard上进行评估时自动生成的。数据集由63个配置组成，每个配置对应一个评估任务。它包含一次运行的数据，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分割始终指向最新结果。此外，还有一个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。README还提供了如何使用Hugging Face datasets库加载数据集的示例，并包含了特定运行的最新结果。

原始信息汇总

数据集概述

该数据集是在评估模型namirocks/mistral-shishya-model-7b-ep3在Open LLM Leaderboard上的运行过程中自动创建的。

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。
• "train"分割始终指向最新的结果。
• 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Open LLM Leaderboard上的聚合指标。

最新结果

以下是2024-01-27T06:00:40.290871运行的最新结果：

python { "all": { "acc": 0.4650442625665404, "acc_stderr": 0.03445571844883224, "acc_norm": 0.4724104681103289, "acc_norm_stderr": 0.03539794125414078, "mc1": 0.22766217870257038, "mc1_stderr": 0.014679255032111068, "mc2": 0.33866365581926344, "mc2_stderr": 0.014776032962007626 }, "harness|arc:challenge|25": { "acc": 0.41467576791808874, "acc_stderr": 0.014397070564409174, "acc_norm": 0.447098976109215, "acc_norm_stderr": 0.014529380160526847 }, "harness|hellaswag|10": { "acc": 0.5856403106950807, "acc_stderr": 0.004916043838455668, "acc_norm": 0.768074088826927, "acc_norm_stderr": 0.004211993665515871 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.37, "acc_stderr": 0.048523658709391, "acc_norm": 0.37, "acc_norm_stderr": 0.048523658709391 }, # 其他任务的结果... }

配置详情

• config_name: harness_arc_challenge_25

  • split: 2024_01_27T06_00_40.290871
    • path: **/details_harness|arc:challenge|25_2024-01-27T06-00-40.290871.parquet
  • split: latest
    • path: **/details_harness|arc:challenge|25_2024-01-27T06-00-40.290871.parquet

• config_name: harness_gsm8k_5

  • split: 2024_01_27T06_00_40.290871
    • path: **/details_harness|gsm8k|5_2024-01-27T06-00-40.290871.parquet
  • split: latest
    • path: **/details_harness|gsm8k|5_2024-01-27T06-00-40.290871.parquet

• config_name: harness_hellaswag_10

  • split: 2024_01_27T06_00_40.290871
    • path: **/details_harness|hellaswag|10_2024-01-27T06-00-40.290871.parquet
  • split: latest
    • path: **/details_harness|hellaswag|10_2024-01-27T06-00-40.290871.parquet

• config_name: harness_hendrycksTest_5

  • split: 2024_01_27T06_00_40.290871
    • path:
      • **/details_harness|hendrycksTest-abstract_algebra|5_2024-01-27T06-00-40.290871.parquet
      • **/details_harness|hendrycksTest-anatomy|5_2024-01-27T06-00-40.290871.parquet
      • **/details_harness|hendrycksTest-astronomy|5_2024-01-27T06-00-40.290871.parquet
      • **/details_harness|hendrycksTest-business_ethics|5_2024-01-27T06-00-40.290871.parquet
      • **/details_harness|hendrycksTest-clinical_knowledge|5_2024-01-27T06-00-40.290871.parquet
      • **/details_harness|hendrycksTest-college_biology|5_2024-01-27T06-00-40.290871.parquet
      • **/details_harness|hendrycksTest-college_chemistry|5_2024-01-27T06-00-40.290871.parquet
      • **/details_harness|hendrycksTest-college_computer_science|5_2024-01-27T06-00-40.290871.parquet
      • **/details_harness|hendrycksTest-college_mathematics|5_2024-01-27T06-00-40.290871.parquet
      • **/details_harness|hendrycksTest-college_medicine|5_2024-01-27T06-00-40.290871.parquet
      • **/details_harness|hendrycksTest-college_physics|5_2024-01-27T06-00-40.290871.parquet
      • **/details_harness|hendrycksTest-computer_security|5_2024-01-27T06-00-40.290871.parquet
      • **/details_harness|hendrycksTest-conceptual_physics|5_2024-01-27T06-00-40.290871.parquet
      • **/details_harness|hendrycksTest-econometrics|5_2024-01-27T06-00-40.290871.parquet
      • **/details_harness|hendrycksTest-electrical_engineering|5_2024-01-27T06-00-40.290871.parquet
      • **/details_harness|hendrycksTest-elementary_mathematics|5_2024-01-27T06-00-40.290871.parquet
      • **/details_harness|hendrycksTest-formal_logic|5_2024-01-27T06-00-40.290871.parquet
      • **/details_harness|hendrycksTest-global_facts|5_2024-01-27T06-00-40.290871.parquet

      其他路径...

}

数据集加载示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_namirocks__mistral-shishya-model-7b-ep3", "harness_winogrande_5", split="train")