open-llm-leaderboard-old/details_yam-peleg__Experiment19-7B

该数据集是在模型 yam-peleg/Experiment19-7B 在 Open LLM Leaderboard 上的评估运行期间自动创建的。数据集由 63 个配置组成，每个配置对应一个被评估的任务。它包含 1 次运行的结果，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。一个额外的配置 results 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 还提供了如何使用 HuggingFace 数据集库加载运行中的详细信息的示例。

原始信息汇总

数据集概述

数据集信息

• 名称: Evaluation run of yam-peleg/Experiment19-7B
• 来源: 自动创建于模型 yam-peleg/Experiment19-7B 在 Open LLM Leaderboard 的评估运行中。
• 组成: 包含63个配置，每个配置对应一个评估任务。
• 创建: 从1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。"train" 分割始终指向最新的结果。
• 额外配置: "results" 配置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_yam-peleg__Experiment19-7B", "harness_winogrande_5", split="train")

最新结果

• 最新结果时间戳: 2024-02-19T18:09:52.338335
• 聚合结果: python { "all": { "acc": 0.6369774461420744, "acc_stderr": 0.03249014878349776, "acc_norm": 0.6366803430351798, "acc_norm_stderr": 0.03316829176635558, "mc1": 0.6046511627906976, "mc1_stderr": 0.017115815632418208, "mc2": 0.7817594211835219, "mc2_stderr": 0.013695289301759589 }, "harness|arc:challenge|25": { "acc": 0.7013651877133106, "acc_stderr": 0.013374078615068749, "acc_norm": 0.7235494880546075, "acc_norm_stderr": 0.013069662474252425 }, "harness|hellaswag|10": { "acc": 0.7001593308105954, "acc_stderr": 0.004572515919210699, "acc_norm": 0.8860784704242183, "acc_norm_stderr": 0.003170666122517656 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.37, "acc_stderr": 0.04852365870939099, "acc_norm": 0.37, "acc_norm_stderr": 0.04852365870939099 }, # 其他任务结果省略... }

配置信息

• 配置名称: harness_arc_challenge_25

  • 数据文件:
    • 分割: 2024_02_19T18_09_52.338335
      • 路径: **/details_harness|arc:challenge|25_2024-02-19T18-09-52.338335.parquet
    • 分割: latest
      • 路径: **/details_harness|arc:challenge|25_2024-02-19T18-09-52.338335.parquet

• 配置名称: harness_gsm8k_5

  • 数据文件:
    • 分割: 2024_02_19T18_09_52.338335
      • 路径: **/details_harness|gsm8k|5_2024-02-19T18-09-52.338335.parquet
    • 分割: latest
      • 路径: **/details_harness|gsm8k|5_2024-02-19T18-09-52.338335.parquet

• 配置名称: harness_hellaswag_10

  • 数据文件:
    • 分割: 2024_02_19T18_09_52.338335
      • 路径: **/details_harness|hellaswag|10_2024-02-19T18-09-52.338335.parquet
    • 分割: latest
      • 路径: **/details_harness|hellaswag|10_2024-02-19T18-09-52.338335.parquet

• 配置名称: harness_hendrycksTest_5

  • 数据文件:
    • 分割: 2024_02_19T18_09_52.338335
      • 路径:
        • **/details_harness|hendrycksTest-abstract_algebra|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-anatomy|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-astronomy|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-business_ethics|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-clinical_knowledge|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-college_biology|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-college_chemistry|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-college_computer_science|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-college_mathematics|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-college_medicine|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-college_physics|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-computer_security|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-conceptual_physics|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-econometrics|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-electrical_engineering|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-elementary_mathematics|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-formal_logic|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-global_facts|5_2024-02-19T18-09-52.338335.parquet
        • **/details_harness|hendrycksTest-high_school_biology|5_2024-02-19T18-09-52.338335.parquet

        • 其他路径省略...