open-llm-leaderboard-old/details_Ppoyaa__Lumina-3.5

该数据集是在Open LLM Leaderboard上对模型Ppoyaa/Lumina-3.5进行评估时自动创建的。数据集由63个配置组成，每个配置对应一个评估任务。数据集由1次运行生成，每次运行在每个配置中表示为特定的拆分，train拆分始终指向最新结果。此外，名为results的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`datasets`库中的`load_dataset`函数加载运行详情的示例。

原始信息汇总

数据集概述

数据集名称

• 名称: Evaluation run of Ppoyaa/Lumina-3.5

数据集组成

• 配置数量: 63个配置
• 每个配置对应: 一个评估任务
• 创建来源: 从1次运行中创建
• 分割命名: 使用运行的时间戳命名
• 训练分割: 总是指向最新的结果

额外配置

• 结果配置: 存储所有运行结果的聚合数据

数据加载示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_Ppoyaa__Lumina-3.5", "harness_winogrande_5", split="train")

最新结果

• 最新结果来源: 2024-04-16T20:59:59.530267
• 详细结果: 包含多个任务的准确率（acc）、标准化准确率（acc_norm）、误差（stderr）等指标

配置详情

• 配置名称: harness_arc_challenge_25

  • 数据文件:
    • 分割: 2024_04_16T20_59_59.530267
      • 路径: **/details_harness|arc:challenge|25_2024-04-16T20-59-59.530267.parquet
    • 分割: latest
      • 路径: **/details_harness|arc:challenge|25_2024-04-16T20-59-59.530267.parquet

• 配置名称: harness_gsm8k_5

  • 数据文件:
    • 分割: 2024_04_16T20_59_59.530267
      • 路径: **/details_harness|gsm8k|5_2024-04-16T20-59-59.530267.parquet
    • 分割: latest
      • 路径: **/details_harness|gsm8k|5_2024-04-16T20-59-59.530267.parquet

• 配置名称: harness_hellaswag_10

  • 数据文件:
    • 分割: 2024_04_16T20_59_59.530267
      • 路径: **/details_harness|hellaswag|10_2024-04-16T20-59-59.530267.parquet
    • 分割: latest
      • 路径: **/details_harness|hellaswag|10_2024-04-16T20-59-59.530267.parquet

• 配置名称: harness_hendrycksTest_5

  • 数据文件:
    • 分割: 2024_04_16T20_59_59.530267
      • 路径:
        • **/details_harness|hendrycksTest-abstract_algebra|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-anatomy|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-astronomy|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-business_ethics|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-clinical_knowledge|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-college_biology|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-college_chemistry|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-college_computer_science|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-college_mathematics|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-college_medicine|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-college_physics|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-computer_security|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-conceptual_physics|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-econometrics|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-electrical_engineering|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-elementary_mathematics|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-formal_logic|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-global_facts|5_2024-04-16T20-59-59.530267.parquet
        • **/details_harness|hendrycksTest-high_school_biology|5_2024-04-16T20-59-59.530267.parquet