open-llm-leaderboard-old/details_amu__orpo-lora-phi2

该数据集是在Open LLM Leaderboard上对模型amu/orpo-lora-phi2进行评估时自动创建的。它由63个配置组成，每个配置对应一个评估任务。数据集是从1次运行中生成的，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"分割始终指向最新的结果。一个额外的配置"results"存储了运行的所有聚合结果，用于计算和显示Open LLM Leaderboard上的聚合指标。可以使用`datasets`库中的`load_dataset`函数加载该数据集。

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 amu/orpo-lora-phi2 在 Open LLM Leaderboard 上的运行过程中自动创建的。

数据集结构

• 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。

数据加载示例

python from datasets import load_dataset data = load_dataset("open-llm-leaderboard/details_amu__orpo-lora-phi2", "harness_winogrande_5", split="train")

最新结果

以下是 2024-03-31T09:12:19.501086 运行的最新结果：

python { "all": { "acc": 0.5821416277179048, "acc_stderr": 0.033770146849493435, "acc_norm": 0.5843594794170675, "acc_norm_stderr": 0.03445851343693101, "mc1": 0.3072215422276622, "mc1_stderr": 0.01615020132132302, "mc2": 0.44496168517736456, "mc2_stderr": 0.014885989610479747 }, "harness|arc:challenge|25": { "acc": 0.5665529010238908, "acc_stderr": 0.014481376224558902, "acc_norm": 0.6032423208191127, "acc_norm_stderr": 0.014296513020180642 }, "harness|hellaswag|10": { "acc": 0.5540728938458475, "acc_stderr": 0.004960516570284905, "acc_norm": 0.7457677753435571, "acc_norm_stderr": 0.004345388614520023 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.3, "acc_stderr": 0.04605661864718381, "acc_norm": 0.3, "acc_norm_stderr": 0.04605661864718381 }, # 其他任务的结果省略... }

配置详情

• config_name: harness_arc_challenge_25

  • 数据文件:
    • 分割: 2024_03_31T09_12_19.501086
      • 路径: **/details_harness|arc:challenge|25_2024-03-31T09-12-19.501086.parquet
    • 分割: latest
      • 路径: **/details_harness|arc:challenge|25_2024-03-31T09-12-19.501086.parquet

• config_name: harness_gsm8k_5

  • 数据文件:
    • 分割: 2024_03_31T09_12_19.501086
      • 路径: **/details_harness|gsm8k|5_2024-03-31T09-12-19.501086.parquet
    • 分割: latest
      • 路径: **/details_harness|gsm8k|5_2024-03-31T09-12-19.501086.parquet

• config_name: harness_hellaswag_10

  • 数据文件:
    • 分割: 2024_03_31T09_12_19.501086
      • 路径: **/details_harness|hellaswag|10_2024-03-31T09-12-19.501086.parquet
    • 分割: latest
      • 路径: **/details_harness|hellaswag|10_2024-03-31T09-12-19.501086.parquet

• config_name: harness_hendrycksTest_5

  • 数据文件:
    • 分割: 2024_03_31T09_12_19.501086
      • 路径:
        • **/details_harness|hendrycksTest-abstract_algebra|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-anatomy|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-astronomy|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-business_ethics|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-clinical_knowledge|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-college_biology|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-college_chemistry|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-college_computer_science|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-college_mathematics|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-college_medicine|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-college_physics|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-computer_security|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-conceptual_physics|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-econometrics|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-electrical_engineering|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-elementary_mathematics|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-formal_logic|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-global_facts|5_2024-03-31T09-12-19.501086.parquet
        • **/details_harness|hendrycksTest-high_school_biology|5_2024-03-31T09-12-19.501086.parquet

        其他路径省略...

}