open-llm-leaderboard-old/details_lex-hue__Delexa-7b-128k

该数据集是在模型 lex-hue/Delexa-7b-128k 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，每个配置对应一个评估任务。数据集是从一次或多次运行中生成的，每次运行都作为每个配置中的一个特定分割存储。train 分割始终指向最新结果。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 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

数据集基本信息

• 数据集名称: Evaluation run of lex-hue/Delexa-7b-128k
• 数据集来源: 自动创建于模型 lex-hue/Delexa-7b-128k 在 Open LLM Leaderboard 的评估运行中。

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。
• 数据集创建: 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。
• 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_lex-hue__Delexa-7b-128k", "harness_winogrande_5", split="train")

最新结果

• 最新结果时间戳: 2024-04-15T21:36:27.350319
• 结果示例: python { "all": { "acc": 0.6496468924225586, "acc_stderr": 0.03207144308728149, "acc_norm": 0.6509773836165138, "acc_norm_stderr": 0.0327145643645653, "mc1": 0.4418604651162791, "mc1_stderr": 0.017384767478986218, "mc2": 0.6217742992950922, "mc2_stderr": 0.015455929661783052 }, "harness|arc:challenge|25": { "acc": 0.6467576791808873, "acc_stderr": 0.013967822714840055, "acc_norm": 0.6825938566552902, "acc_norm_stderr": 0.013602239088038169 }, "harness|hellaswag|10": { "acc": 0.6879107747460665, "acc_stderr": 0.004623990785158488, "acc_norm": 0.8650667197769368, "acc_norm_stderr": 0.003409540533249841 }, "harness|hendrycksTest-abstract_algebra|5": { "acc": 0.33, "acc_stderr": 0.04725815626252606, "acc_norm": 0.33, "acc_norm_stderr": 0.04725815626252606 },

  其他任务结果省略...

}

配置详情

• 配置名称: harness_arc_challenge_25

  • 数据文件:
    • 分割: 2024_04_15T21_36_27.350319
      • 路径: **/details_harness|arc:challenge|25_2024-04-15T21-36-27.350319.parquet
    • 分割: latest
      • 路径: **/details_harness|arc:challenge|25_2024-04-15T21-36-27.350319.parquet

• 配置名称: harness_gsm8k_5

  • 数据文件:
    • 分割: 2024_04_15T21_36_27.350319
      • 路径: **/details_harness|gsm8k|5_2024-04-15T21-36-27.350319.parquet
    • 分割: latest
      • 路径: **/details_harness|gsm8k|5_2024-04-15T21-36-27.350319.parquet

• 配置名称: harness_hellaswag_10

  • 数据文件:
    • 分割: 2024_04_15T21_36_27.350319
      • 路径: **/details_harness|hellaswag|10_2024-04-15T21-36-27.350319.parquet
    • 分割: latest
      • 路径: **/details_harness|hellaswag|10_2024-04-15T21-36-27.350319.parquet

• 配置名称: harness_hendrycksTest_5

  • 数据文件:
    • 分割: 2024_04_15T21_36_27.350319
      • 路径:
        • **/details_harness|hendrycksTest-abstract_algebra|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-anatomy|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-astronomy|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-business_ethics|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-clinical_knowledge|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-college_biology|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-college_chemistry|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-college_computer_science|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-college_mathematics|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-college_medicine|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-college_physics|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-computer_security|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-conceptual_physics|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-econometrics|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-electrical_engineering|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-elementary_mathematics|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-formal_logic|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-global_facts|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-high_school_biology|5_2024-04-15T21-36-27.350319.parquet
        • **/details_harness|hendrycksTest-high_school_chemistry|5_2024-04-15T21-36-27.350319.parquet

        其他文件路径省略...

}